Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The evolution of bird pollination in Macaronesian Lotus section Rhyncholotus (Leguminosae) Ojeda Alayon, Isidro 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2011_spring_ojedaalayon_isidro.pdf [ 4.79MB ]
Metadata
JSON: 24-1.0071648.json
JSON-LD: 24-1.0071648-ld.json
RDF/XML (Pretty): 24-1.0071648-rdf.xml
RDF/JSON: 24-1.0071648-rdf.json
Turtle: 24-1.0071648-turtle.txt
N-Triples: 24-1.0071648-rdf-ntriples.txt
Original Record: 24-1.0071648-source.json
Full Text
24-1.0071648-fulltext.txt
Citation
24-1.0071648.ris

Full Text

THE EVOLUTION OF BIRD POLLINATION IN MACARONESIAN LOTUS SECTION RHYNCHOLOTUS (LEGUMINOSAE)  by  Isidro Ojeda Alayón  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  The Faculty of Graduate Studies  (Botany)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  March, 2011  © Isidro Ojeda Alayón, 2011  ABSTRACT  In order to understand the evolutionary transition from bee pollination (melittophily) to bird pollination (ornithophily), I studied a group of Lotus from Macaronesia. First, I provided a combined phylogenetic framework using nuclear and plastid genes, where I showed that the morphological features adapted to opportunistic passerine birds in Tenerife and La Palma are derived and evolved recently within the last 1.2 Ma in four species. I also identified Lotus sessilifolius as the most likely closely related species with melittophily. I showed that L. sessilifolius and the clade where this syndrome evolved had a pre-adaptation to produce a color change to red flowers (and the associated anthocyanidin pigment, cyanidin) as a possible strategy to increase bee foraging efficiency. The transition from yellow to red flowers in this group required only a redirection in the flux of pigment production and a modification in the proportions of flavonols and anthocyanidins, especially within the cyanidin branch. I also found that petal micromorphology is highly modified between the two syndromes. I found that ornithophilous flowers lack the typical papillose conical cells, which are distributed in the exposed sides of the petals in bee-pollinated flowers. This reduction of conical cells is associated with an early down-regulation of a dorsal identity gene in legumes, LjCYC2. Bird-pollinated flowers also have a higher proportion of tabular rugose cells in all three types of petals in comparison with the bee-pollinated species. The increase of this epidermal type is associated with an up-regulation of LjCYC3, a lateral petal identity gene, during early stages of flower development in the dorsal and lateral petals. Lotus sessilifolius also seems to have this early expression in comparison with other bee-pollinated species. All this evidence suggests that the transition from bee to bird pollination in this group required only heterochronic modifications of the genes involved in flower color and petal micromorphology. It seems that ornithophily in Lotus evolved within a group which has at least two pre-adaptations, production of pigments required for red colors and an increase in the amount of tabular rugose cells, which likely facilitated the evolution of phenotypes associated with this pollination syndrome in the Canary Islands.  ii  TABLE OF CONTENTS Abstract .................................................................................................................................... ii Table of contents ..................................................................................................................... iii List of tables........................................................................................................................... viii List of figures ........................................................................................................................... xi Acknowledgements ............................................................................................................... xvii Dedication ............................................................................................................................ xviii Co-authorship statement ........................................................................................................ xx 1. Introduction .......................................................................................................................... 1 1.1 Bird pollination .................................................................................................................... 1 1.2 Macaronesian Lotus ............................................................................................................. 2 1.3 Molecular phylogenetics and DNA barcoding of Macaronesian Lotus .................................. 3 1.4 Flower colour and petal micromorpholgy modifications during the transition from bee to bird pollination in Macaronesian Lotus ......................................................................................... 4 1.5 Bibliography......................................................................................................................... 5 2. Bird-pollinated flowers in an evolutionary and molecular context .................................... 9 2.1  Pollination syndromes....................................................................................................... 9  2.2  The syndrome of ornithophily ......................................................................................... 10  2.2.1 The birds...................................................................................................................... 10 2.2.2 Ecology of bird pollination .......................................................................................... 11 2.2.3 Perching vs. hovering................................................................................................... 12 2.2.4 Hermit vs. non-hermit hummingbirds........................................................................... 13 2.2.5 Major groups of angiosperms with bird pollination ...................................................... 14 2.3 The floral phenotype ........................................................................................................... 16 2.3.1 Types of floral adaptation ............................................................................................ 16 2.3.2 Why red? ..................................................................................................................... 17 2.3.3 Flower colour and its perception .................................................................................. 18 2.3.4 Nectar .......................................................................................................................... 20 2.3.5 Corolla morphology in bird-pollinated flowers ............................................................. 22 2.3.6 Other characters associated with bird pollination.......................................................... 23 2.4 The molecular basis for the evolution of ornithophily ......................................................... 26 2.4.1 Some model systems ..................................................................................................... 26 2.6 Bibliography....................................................................................................................... 32 iii  3. Evolution of petal epidermal micromorphology in Leguminosae and its use as a marker of petal identity .................................................................................................................. 40 3.1 Introduction ........................................................................................................................ 40 3.1.1 Objectives of the study.................................................................................................. 42 3.2 Materials and methods ........................................................................................................ 43 3.2.1 Taxon sampling ............................................................................................................ 43 3 2.2 Microscopy and cell type classification ......................................................................... 44 3.2.3 Reconstruction of evolutionary changes ........................................................................ 46 3.3 Results................................................................................................................................ 46 3.3.1 Epidermal types within Leguminosae ............................................................................ 46 3.3.2 Strong micromorphological variation in petals is exclusive to papilionoid legumes ....... 47 3.3.4 Loss of micromorphological variation in Indigofera, Amorpheae and IRLC .................. 48 3.3.5 Zygomorphy in caesalpinioids is not associated with strong micromorphological variation ............................................................................................................................................ 49 3.3.6 The occurrence of papillose cells in the Leguminosae ................................................... 50 3.4 Discussion .......................................................................................................................... 51 3.4.1 Functional significance of papillose cell types............................................................... 51 3.4.2 Lack of micromorphological variation in the Caesalpinioideae ..................................... 51 3.4.3 Mixing of epidermal types in the same petal surface ..................................................... 52 3.4.4 Contrasting loss of micromorphological variation within flowers of the Amorpheae and IRLC ................................................................................................................................... 53 3.4.5 Genetic control of petal micromorphology and petal identity ........................................ 54 3.4.6 Evolution of petal micromorphology ............................................................................. 56 3.5 Bibliography....................................................................................................................... 77 4. The origin of bird pollination in Macaronesian Lotus (Loteae, Leguminosae) ................ 84 4.1 Introduction ........................................................................................................................ 84 4.1.1 Objectives of the study.................................................................................................. 88 4.2 Materials and methods ........................................................................................................ 88 4.2.1 Taxon sampling ............................................................................................................ 88 4 2.2 DNA Extraction and sequence data analysis .................................................................. 90 4.2.3 Nuclear regions ............................................................................................................. 91 4.2.4 Plastid regions .............................................................................................................. 91 4.2.5 Phylogenetic analyses and character ancestral reconstruction ........................................ 92 iv  4.2.6 Dating the origin of bird pollination .............................................................................. 93 4.3 Results................................................................................................................................ 94 4.3.1 Major clades of Pedrosia s.l. in Macaronesia ................................................................ 94 4.3.2 Recovery of the sister group of Rhyncholotus using a six gene region analysis with the inclusion of CYCLOIDEA homologues ................................................................................ 96 4.3.3 Dating the origin of bird-pollinated Lotus in the Canary Islands ....................................... 97 4.4 Discussion .......................................................................................................................... 98 4.4.1. Closest relative of the four bird-pollinated species ....................................................... 98 4.4.2. How old are bird-pollinated Lotus from Macaronesia? ............................................... 100 4.4.3. The availability of new niches and the evolution of ornithophily in Lotus .................. 103 4.5 Bibliography..................................................................................................................... 128 5. Barcoding plant island radiations and its applicability in species recognition and conservation in Macaronesian Lotus section Pedrosia and Rhyncholotus (Loteae, Leguminosae) ................................................................................................................... 137 5.1 Introduction ...................................................................................................................... 137 5.1.1 Objectives of the study................................................................................................ 140 5.2 Materials and methods ...................................................................................................... 140 5.2.1 Taxon sampling .......................................................................................................... 140 5 2.2 Selecting barcode regions ........................................................................................... 141 5.2.3 Molecular analysis ...................................................................................................... 141 5.2.4 Assessment of the barcode regions .............................................................................. 142 5.3 Results.............................................................................................................................. 143 5.3.1 Universality and sequence quality ............................................................................... 143 5.3.2 Species discrimination of the plastid regions ............................................................... 143 5.3.3 ITS as a barcode in Lotus ............................................................................................ 144 5.4 Discussion ........................................................................................................................ 144 5.4.1 ITS vs. plastid data ..................................................................................................... 147 5.5 Bibliography..................................................................................................................... 161 6. Temporal, but not spatial, changes in expression patterns of floral identity genes are associated with the evolution of bird pollination in Lotus (Leguminosae) ..................... 167 6.1 Introduction ...................................................................................................................... 167 6.1.1 Objectives of the study................................................................................................ 169 6.2 Materials and methods ...................................................................................................... 170 6.2.1 Plant material and growth conditions .......................................................................... 170 v  6.2.2 Flower developmental stages and petal growth ........................................................... 170 6.2.3 Scanning electron microscopy (SEM) and light microscopy ........................................ 171 6.2.4 Semi-quantitative reverse RT-PCR ............................................................................. 172 6.3 Results.............................................................................................................................. 173 6.3.1 Petal micromorphology of flowers at anthesis within Loteae and the lateralization of petals in section Rhyncholotus and Pedrosia ...................................................................... 173 6.3.2 Expression patterns of CYCLOIDEA homologues during flower development ............ 174 6.4 Discussion ........................................................................................................................ 175 6.4.1 Lateralization of the dorsal petal in the bird-pollinated species and the adaptive value of papillose conical cells in melittophilous species ................................................................. 175 6.4.2 Late expression of LjCYC2 is associated with modification of epidermal types and the evolution of ornithophily ................................................................................................... 177 6.4.3 Early expression of LjCYC3 is associated with a lateralization of the flower and suggests a molecular pre-adaptation to bird pollination ....................................................................... 179 6.5 Bibliography..................................................................................................................... 193 7. Biochemical evolution of bird-pollinated flowers in Macaronesian Lotus (Leguminosae) .......................................................................................................................................... 198 7.1 Introduction ...................................................................................................................... 198 7.1.1 Objectives of the study................................................................................................ 201 7.2 Materials and methods ...................................................................................................... 201 7.2.1 Pigment extraction and composition analysis .............................................................. 201 7.2.2 Measurement of petal reflectance ................................................................................ 202 7.2.3 Reconstruction of flower color change evolution ........................................................ 203 7.2.4 Gene expression of dihydroflavonol-4-reductase (DFR), Anthocyanidin synthase (ANS) and O-methyl transferase (LjOMT) .................................................................................... 204 7.3 Results.............................................................................................................................. 205 7.3.1 Pigment flower composition in Lotus .......................................................................... 205 7.3.2 Flower reflectance and color change perception .......................................................... 208 7.3.3 Gene expression comparisons during flower color modification .................................. 208 7.4 Discussion ........................................................................................................................ 209 7.4.1 The changing balance of anthocyanin and flavonol pigment composition in bee and bird pollination in Lotus ............................................................................................................ 209 7.4.2 Post-pollination floral colour change as a mechanism to promote foraging efficiency . 211 7.4.3 Late-anthesis flower colour changes as a possible preadaptation in the evolution of the bird pollination syndrome in Lotus ..................................................................................... 212 vi  7.4.4 Flower color change after anthesis as a possible facilitating factor in the evolution of bird pollination in other groups ................................................................................................. 212 7.4.5 Red flowers in bird-pollinated Lotus may avoid bee visits ........................................... 213 7.4.6 Down-regulation of O-methyltransferase (LjOMT) in bird-pollinated Lotus ................ 214 7.4.7 Possible specialization of dihydroflavonol 4-reductase (LjDFR) copies in Lotus ......... 214 7.5 Bibliography..................................................................................................................... 228 8. Conclusion ........................................................................................................................ 235 8.1 Conclusions and future directions ..................................................................................... 235 8.2 Bibliography..................................................................................................................... 243  vii  LIST OF TABLES Table 3.1 List of species sampled during this study. Clades are recognized following Wojciechowski et al., 2004; Lavin et al., 2005; Lewis et al., 2005. Tribal classification following Lewis et al., 2005. *= species analyzed from fresh flowers preserved in ethanol using a light microscope, += species analyzed from flowers re-dehydrated from voucher specimens and preserved in ethanol using a light microscope. Other species were studied using the SEM and fresh material. UBCBG= UBC Botanical Garden, FTBG= Fairchild Tropical Botanical Garden, MBG= Montgomery Botanical Garden, QEG= Queen Elizabeth Garden, Vancouver, JBRCICY= Jardín Botánico Regional del CICY, CICY= Centro de Investigación Científica de Yucatán, Mexico. JAO= Jardín de Aclimatación de la Orotova, Spain.. ........... 54 Table 3.2 Distribution of the major epidermal types in each of the three types of petal in the Leguminosae. PCS= papillose conical cells, PKR= papillose knobby with rugose sculpture, PLS= papillose lobular cells, TRG= tabular rugose with a granular sculpture, TRS= tabular rugose striate, (TRSi TRSii TRSiii represents minor variations within this epidermal type that allow petal identification within the same species) and TFS= tabular flat longitudinally striate. More than one major epidermal type with more or less equal distribution are separated by a slash. Cell types in bold indicates the side more differentiated in each type of petal. If both sides are bold, then more or less the same level of differentiation is implied. PLS is a rare type that is never characteristic of whole petals and its presence is therefore only noted by the symbol ‡ (it is found in some Lathyrus species where it is restricted to the margin of the dorsal petal). s= stomata, t= trichomes, st= trichomes and stomata. *= flowers preserved in ethanol and analyzed using a light microscope, += flowers re-dehydrated from voucher specimens, preserved in ethanol and analyzed using a light microscope. Other species were studied using the SEM and fresh material. -= petals absent.................................................. 60 Table 3.3  Classification of the epidermal types observed in Leguminosae ............................. 65  Table 3.4 Distribution of the major epidermal cell types within sampled Papilionoideae. PCS= papillose conical cells, PKR = papillose knobby cells, TRS= tabular rugose cells with striations, TFS= tabular flat cells with striations ................................................................... 65 Table 4.1 Samples included in the phylogenetic analysis of the two data sets with 21 samples only (*) and with the 54 samples within Pedrosia s. l. JAO= Jardín de Aclimatacion de la Orotava, JBCVC= Jardín Botánico Canario ―Viera y Clavijo‖, UBC= University of British Columbia. T= Tenerife, GC= Gran Canaria, G= La Gomera, P=La Palma, H=El Hierro, CV=Cape Verde, M=Madeira ............................................................................................ 105 Table 4.2 Species included in the phylogenetic analysis with the nuclear ribosomal ITS region only. GenBank sequences and new sequences generated in this analysis. I excluded six sequences from GenBank (L. dumetorum AY294294, L. campylocladus AF450196, L. creticus AF450192, L. arinagensis FJ411112, L. loweanus FJ 411117, L. mascaensis FJ411118 ) due to the ambiguities in the sequences and/or suspected misidentification. JAO= Jardín de Aclimatación de la Orotava, JBCVC= Jardín Botánico Canario ―Viera y Clavijo‖, UBC= University of British Columbia. T= Tenerife, GC= Gran Canaria, G= La Gomera, P=La Palma, H=El Hierro, CV=Cape Verde, M=Madeira ....................................................................... 107 viii  Table 4.3  Nuclear and plastid regions used in the three data sets analyzed in this study ....... 111  Table 4.4 Information of the four nuclear and plastid regions used in the phylogenetic reconstruction of Pedrosia s.l. with the ITS only including GenBank sequencesa, excluding ITS sequences from GenBankb and with the ITS data set combined with the plastid regionsc. *Using maximum parsimony ............................................................................................. 112 Table 4.5 Gene regions used with the 21 sample data set used to identify the closest relative species of the four rhyncholotus species within clade B. Variability of each region when analyzed separate and in combination ................................................................................ 113 Table 4.6 Date of origin (MRCA) in Ma of various clades based on two data sets. Values obtained for each clade when rhyncholotus constrained or unconstrained to be monophyletic .......................................................................................................................................... 113 Table 5.1 Macaronesian Lotus species considered under different levels of threat. According to Red List of Spanish Vascular Flora based on the IUCN Red Data Book (IUCN) (VV. AA., 2000), the Atlas of Endangered Spanish Vascular Flora (AESVF) (Bañares et al., 2004), and the ranking according to the Top 100 endangered species of Macaronesia (Martin et al., 2008). Numbers indicate their rank under the Top 100 list, - = not considered within the 100 most endangered species. CR= critically endangered, EN= endangered, VU= vulnerable ........... 148 Table 5.2 Species from the sections Pedrosia and Rhyncholotus sampled in this analysis. Distribution: G= La Gomera, P=La Palma, T=Tenerife, GC= Gran Canaria, CV= Cape Verde, M= Madeira, H= Hierro, L= Lanzarote, F= Fuerteventura. UBC= University of British Columbia, LBCVC= Jardín Botánico Canario Viera y Clavijo, JAO= Jardín de Aclimatación de la Orotava ..................................................................................................................... 149 Table 5.3 Sections Pedrosia and Rhyncholotus and their informal classification based on morphological features. *According to Sandral et al (2006). ¶Species not sampled in this analysis .............................................................................................................................. 152 Table 5.4 Performance of the five plastid regions and the nuclear ribosomal ITS tested separate and in two-pair combinations. A= including all accessions, B= excluding accessions with missing sequences in two-pair combinations. * Informal sections according to Sandral et al., (2006) ................................................................................................................................ 153 Table 5.5 Nuclear and plastid gene regions tested in this analysis with their specific primers and performance ....................................................................................................................... 154 Table 6.1  List of samples collected for the analysis of epidermal types in Loteae................. 181  Table 6.2 Distribution of epidermal types on species analyzed. PCS= papillose conical cells, TRS= tabular rugose cells with striations, TFS= tabular flat cells. Epidermal types separated by a dash indicate that two epidermal types where observed on the same petal, s= stomata, t= trichomes ........................................................................................................................... 183 Table 6.3 Classification of the levels of lateralization (measured by the presence of TRS) observed in Rhyncholotus and Pedrosia species according with the distribution of epidermal types and trichomes. * Species reported with trichomes on the dorsal petal (Sandral et al., ix  2006), but not confirmed in this study. L. chazalei and L. loweanus were not analyzed. The specimen of L. assakensis analyzed did not have trichomes ............................................... 185 Table 7.1 List of species in section Pedrosia (including bird-pollinated members of the rhyncholotus group) analyzed for pigment composition ..................................................... 217 Table 7.2 Lotus species from section Pedrosia (including bird-pollinated members of the rhyncholotus group) measured for petal reflectance ........................................................... 218 Table 7.3 Relative amounts of anthocyanidins and flavonols identified in Lotus section Pedrosia and Rhyncholotus. Relative amount was estimated based on the reflectance peak from each pigment ............................................................................................................. 219 Table 7.4 Classification of flower reflectance of bee and bird-pollinated flowers according to human and bee perception. Bold numbers indicate pigment concentration after flower colour change ............................................................................................................................... 220  x  LIST OF FIGURES Figure 2.1 Forms of bird-pollinated flowers. (A) Strelitzia reginae. (B) Erythrina suberosa, (C) Babiana ringens with sterile inflorescences for perching birds (arrow). (D) Cadia purpurea, a member of the Genistioids with radial symmetry with nectar globes (arrow). (E) Ipomopsis aggregata. (F) Phygelius capensis. (G) Psittacanthus sp. (H) Fritillaria suberosa with nectar globes (arrow)...................................................................................................................... 27 Figure 2.2 Approximate world distributions of the three main families of flower visiting birds: hummingbirds (Trochilidae), sunbirds (Nectariniidae), and honey-eaters (Meliphagidae)..... 28 Figure 2.3 Flowers of Lotus species. Bird-pollinated species of (A) Lotus berthelotii Masf. and (B) L. maculatus Breitfeld, two members of the subgenus Pedrosia s.l. (D) The model legume L. japonicus GIFU B129 and (E) L. arenarius Brot. a closely related species of the birdpollinated species within the subgenus Pedrosia s.l. (C, F) Diagrammatic representation of the hypothetical mechanism by which a bird seeks nectar in flowers of L. berthelotii. Arrows indicate the direction in which the dorsal petal is pushed. a, Dorsal petal; b, lateral petal; and c, ventral petal. (C, F) Modified from Olesen (1985) ............................................................... 29 Figure 3.1 Diversity of flower symmetry in the Leguminosae. (A) Zygomorphic flowers of Cercis canadensis (Caesalpinioideae). (B) Flowers of Bauhinia tomentosa (Caesalpinioideae). (C) Radially symmetric flower of Brownea capitella (Caesalpinioideae). (D) Zygomorphic flowers of Tara cacalao (Caesalpinioideae). (E) Asymmetric enantiostylous flower in Cassia emarginata (Caesalpinioideae). (F) Zygomorphic flower with the dorsal petal differentiated with respect of the other petals in Caesalpinia pulcherrima (Caesalpinioideae). (G) Radial flower with reduced petals in Inga paterno (Mimosoideae). (H) Zygomorphic flowers in Lespedeza thunbergii (Papilionoideae) (with the ventral petal more exposed), (I) Clitoria ternatea (Papilionoideae) (the dorsal petal is pointing downwards), (J) Sesbania punicea (Papilionoideae), and (K) Lathyrus sylvestris (Papilionoideae) (ventral petals enclosed) ...... 66 Figure 3.2 Four axes of variation considered in the study of the epidermal types and their distribution on each petal. A) Lotus corniculatus, three types of petals in zygomorphic papilionoid flowers, a) dorsal, b) lateral and c) ventral, within an adaxial-abaxial axis within the flower. B) Abaxial-adaxial surface within the petal. The abaxial side is exposed in lateral (b) and ventral (c) petals in most papilionoids. However, the adaxial surface is exposed in most standard petals (a). C) Senna corymbosa with three types of petals, (a) dorsal, (b) lateral and c) ventral within an adaxial-abaxial axis within the flower. Further axes are: proximal-distal and medio-lateral axes within D) the dorsal petal, E) lateral petal and F) ventral petal in Lotus corniculatus. The base of the dorsal petal (claw) in L. corniculatus has been separated from the rest of the petal .................................................................................................................... 67 Figure 3.3 Classification of the major epidermal types in Leguminosae. (A, G, M) tabular rugose cells with granulose sculpture (TRG) in Calliandra haematocephala (Mimosoideae). (B and H) tabular rugose cells with striation (TRS) in Wisteria sinensis (Papilionoideae) and (N) in Lotus japonicus (Papilionoideae). (C and I) tabular flat cells with striations (TFS) in W. sinensis and (O) in Lotus japonicus. (D, J, P) papillose conical cells (PCS) in the dorsal petal of Lotus japonicus. (E, K, Q) papillose knobby cells (PKR) in the dorsal and lateral petals of xi  Robinia pseudoacacia (Papilionoideae). (F and L) papillose lobular cells (PLS) in the dorsal petal of Lathyrus venetus and (R) in Lathyrus sylvestris (Papilionoideae). This latter epidermal type was only observed in these two species. All images correspond to the adaxial side of the petal. Scale bar 50 µm (A-F), 100 µm (G-L) and 20 µm (M-R) ............................................ 68 Figure 3.4 Minor epidermal types within the tabular rugose cells with striations (TRS). Minor epidermal variants within TRS are designated as i, ii and iii. (A-C) variation among minor epidermal types in Pisum sativum (Papilionoideae) enables each petal type to be distinguished. Variation within minor epidermal types distinguished the ventral but does not distinguish between the lateral and dorsal petals in (D-F) Vicia hirsuta (Papilionoideae), (G-I) Trifolium repens (Papilionoideae). Variation among minor epidermal types does not allow clear characterization of the three petal types in (J-L) Melilotus officinalis (Papilionoideae) and (MO) Cercis canadensis (Caesalpinioideae). All illustrations correspond to the adaxial side of the petal, except H, I, K, L, N and O, which correspond to the abaxial side. All scale bars 50 µm . ....................................................................................................................................................69 Figure 3.5 Distribution of the epidermal types along the dorsiventral (adaxial-abaxial) axis within the flower. A) micromorphological variation in Lotus burttii (as in almost all Loteae) (Papilionoideae), B) lack of micromorphological variation of major epidermal types in Cassia roxburghii (Caesalpinioideae) with only PKR, C) Cassia emarginata (Caesalpinioideae) with only TRS, and D) Senna alata (Caesalpinioideae) with only PCS on all petals. All petals have the adaxial side shown, except L. burttii, where the abaxial side is presented on lateral and ventral petals and Senna alata where images shown the abaxial side. Scale bars A) 20 µm, B) 50 µm, C) 100 µm, D) 50 µm............................................................................................... 70 Figure 3.6 Schematic representation of the phylogenetic relationships within Leguminosae showing the typical distribution of the major epidermal types observed. This figure is intended to summarise the main patterns but it should be noted that rare variant patterns may occur in clades as well as those listed. The general epidermal surface observed within the flower is given in relation to the dorsiventral axis within the flower, using the representation: dorsal/lateral/ventral petal. Clades with an asterisk contain lineages with loss of papillose cells (PCS and PKR). The numbers following the asterisk indicate the number of losses under parsimony and ML, respectively. -- lack of this type of petal (Tree according to Wojciechowski et al., 2004; Lavin et al., 2005; Lewis et al., 2005) ...................................... 71 Figure 3.7 Distribution of epidermal types along a proximo-distal axis within the same petal. (AE) Lupinus littoralis with transitions from poorly differentiated cells at the base of the adaxial side in the dorsal petal (C), to papillose cells (PCS) on the central part and apex of the petal (A,B). (F-J) the abaxial side of the dorsal petal in Lotus japonicus with transitions from poorly differentiated cells at the base of the petal (H) to PCS in the central and distal regions (G, H). A photograph of the standard petal of Lotus japonicus is shown besides the images with the claw separated and shown below. (K-P) the adaxial side of the dorsal petal in Lathyrus sylvestris has a transition zone from TRS to PLS on the borders of the petal with a transition zone where cells have a mixture of morphological features of both epidermal types. (Q-V) the abaxial side of the lateral petal in Lotus japonicus where TRS is mainly observed at the base and in the central part of the petal and there is a transition zone from TRS to PCS where the xii  cells have a mixture of both epidermal types. Scale bars 500 µm in B; 200 µm in G; 100 µm in A, C, F, J-M; 50 µm in D, H, I, O, P, Q-S; 20 µm in E, N, T-V ............................................ 72 Figure 4.1 Geographical distribution of Lotus section Pedrosia and Rhyncholotus in the Macaronesian region, including five Atlantic volcanic archipelagos (Madeira, Azores, the Salvage islands, the Canary Islands and Cape Verde islands), Europe and Africa. Each archipelago with the No. of species/ No. of endemic species. The age of current above-sea level for each island according to Carracedo et al. (2002). The phytogeographic region of Macaronesia in dashed lines, including a portion of Africa mainland denominated as the ―Macaronesian enclave‖ (Kim et al. 2008) ......................................................................... 114 Figure 4.2 Geographical distribution of Lotus sections Pedrosia and Rhyncholotus in the Canary Island archipelago, with the oldest ages of the subaerial volcanism of each island according to Carracedo et al. (2002). No. of species/No. of endemic species on each island ................... 115 Figure 4.3 Strict consensus tree of the phylogenetic relationships using parsimony within Pedrosia s.l. using ITS DNA sequences. The four clades recovered are labelled (clade A-D). * indicate species groups previously recognized in Lotus according to Sandral et al (2006). Values above branches represent bootstrap values/Bayesian support .................................. 116 Figure 4.4 Strict consensus tree based on a combined analysis of three plastid regions (trnHpsbA, matK and CYB6) using maximum parsimony. Values above branches indicate bootstrap values. Clades A, B and C from the ITS analysis are shown superimposed on the tree ....... 118 Figure 4.5 Majority tree recovered from a combined analysis of one nuclear (ITS) and three plastid gene regions (matK, trnH-psbA and CYB6) using maximum parsimony. Branches with an arrow indicate clades not observed in the strict consensus. Values above branches represent bootstrap from MP/posterior probabilities support ............................................................. 119 Figure 4.6A Maximum parsimony strict consensus tree based on four nuclear regions (ITS, LjCYC 1, 2 and 3) and two plastid regions (matK and trnH-psbA) using maximum parsimony. Values above branches indicate bootstrap values/ Bayesian support. Values below 50 are not indicated. Arrows indicate clades recovered in a MP, Bayesian and ML analyses using this data set............................................................................................................................... 121 Figure 4.6B Maximum likelihood tree using a data set of six genes and 21 samples with Garli 122 Figure 4.7 Flower and leaflet morphology in L. argyrodes (A-C), L. mascaensis (D-F) and L. sessilifolius var. pentaphyllus (G-I). ................................................................................... 123 Figure 4.8 Flower and leaflet morphology in the four bird-pollinated species. Both groups are differentiated by minor morphological features including the number of flowers per inflorescence, the position and orientation of the dorsal petal and the position of the lateral petal. In the species from La Palma the lateral petal is fused at the tip and covers the tip of the ventral petal. In contrast, species from Tenerife have a ventral petal more exposed....……...124 Figure 4.9 Chronogram obtained for the evolution of bird pollination in Macaronesian Lotus under a Bayesian relaxed clock uncorrelated clock model using Beast and applied to the combined data set of 52 samples and using a data set of four gene regions (ITS, matK, trnHpsbA and CYB6). Upper limits of the ages of La Palma (1.77 Ma), El Hierro (1.12 Ma), and xiii  Fuerteventura (21 Ma) were used as calibration points (black circles). Bird-pollinated species are shown in red branches. Ages estimates with their 95% credibility intervals are shown on nodes. Values on grey squares represent bootstrap values from MP/posterior probabilities inferred from the Bayesian inference. Major geological events between the Miocene and Pleistocene are indicated with arrows at the bottom. Clades named after the groups recovered with ITS and the four gene data analyses ........................................................................... 125 Figure 4.10 Chronogram of the rhyncholotus group using a six gene data set and analyzed with Beast using El Hierro (1.12 Ma), La Palma (1.77 Ma) and the age of Gran Canaria (14.5 Ma) as upper age estimates (black circles). Node ages are indicated above branches with 95% HDP. .......................................................................................................................................... 114 Figure 5.1 NJ tree generated with the combination of the CBOL recommended 2-locus, matK+rbcL. Gray squares represent species with more than one sample and species in a square represent species with a single accession. Branches with a black square represent informal taxonomic groups identified. Species in bold belong to section Rhyncholotus while species not in bold are included within section Pedrosia. *endangered species identified ..................... 155 Figure 5.2 NJ tree generated with the combination of all five plastid regions. Gray squares represent species with more than one sample and species in a square with a single accession. Branches with a black square represent informal taxonomic groups identified. Species in bold belong to section Rhyncholotus while species not in bold are included within section Pedrosia. * endangered species identified .......................................................................................... 157 Figure 5.3 NJ tree generated with the combination of all six regions tested (rbcL, matK, trnHpsbA, rpoC1, rpoB and ITS). Gray squares represent species with more than one sample and species in a square represent species with a single accession. Branches with a black square represent informal taxonomic groups identified. Species in bold belong to section Rhyncholotus while species not in bold are included within section Pedrosia.* endangered species identified......................................................................................................................159 Figure 6.1 Developmental stages of flower in Lotus japonicus.............................................. 186 Figure 6.2 Members of Lotus section Rhyncholotus with an ornithophilous pollination syndrome. A, Lotus maculatus, B, L. pyranthus, C, L. berthelotii, and D, L. eremiticus; and four representative species of section Pedrosia, E, L. arenarius, F, L. latifolius, G, L. jacobaeus and H, L. sessilifolius subsp. sessilifolius with a bee-pollination syndrome. a= dorsal petal, b= lateral petal and c= ventral petal. Photo credits: A, C, E, H from I. Ojeda and B, D, F, and G from F. Oliva-Tejera .......................................................................................................... 187 Figure 6.3 Zygomorphic flowers of A, Lotus japonicus. B, these flowers have three types of petals, a) dorsal petal with a bilateral symmetry, b) two asymmetrical lateral petals, and c) two asymmetrical ventral petals. All petals were separated and flattened. The base of the dorsal petals in both species was separated from the rest of the petal. C, Lotus berthelotii with a birdpollination syndrome. D, size comparison between L. berthelotii and L. japonicus. E, hypothetical mechanism by which a bird seeks for nectar and pollen is placed either on the top of the head or in the throat (according to Olsen, 1985). E, Bombus canariensis canariensis one xiv  of the main insect pollinators in the Canary Islands foraging on L. hillebrandii from El Hierro .......................................................................................................................................... 188 Figure 6.4 Major epidermal types recorded in Loteae. A and B, papillose conical cells (PCS), C and D, tabular rugose cells with striations (TRS) and E an F, tabular flat cells with striation (TFS) in Lotus japonicus. G, non-differentiated cells with trichomes , H, tabular rugose cells with striations, I, papillose conical cells with striations, H, J-M tabular rugose cells with striations in L. berthelotii (TRS). Arrows indicate the localization of papillose conical cells in the lateral petals. Scale bars: 100 μm, G and L. 50 μm= A, H, C, D, F, K and M. 25 μm= I and J .................................................................................................................................. 189 Figure 6.5 Expression patterns of LjCYC1, LjCYC2 and LjCYC3 in A) Lotus japonicus, B) L. filicaulis both from section Lotus, C) L. sessilifolius from section Pedrosia. All these three species have a flower morphology adapted to bee pollination and have similar petal size and shape. D) L. berthelotii and E) L. maculatus from section Rhyncholotus have a flower morphology adapted to ornithophily by opportunistic passerines. The base of the dorsal petal has been detached from the rest of the petal. dp= dorsal petal, lp= lateral petal and vp=ventral petal ....................................................... 190 Figure 6.6 Petal growth measured as the ratio length/width during different stages of flower development in A) L. japonicus, B) L. sessilifolius and C) L. berthelotii. Mean values with standard deviations (n=3-5 flowers on each developmental stage). ..................................... 192 Figure 7.1 Red-orange flowers in the bird-pollinated species of the ―rhyncholotus group‖, (a) Lotus berthelotii and (b) L. pyranthus. Yellow flowers of bee-pollinated species that do not modify flower color after anthesis (c) L. campylocladus, and late-anthetic flowers after color change in (d) L. glaucus, (e) L. eriosolen, (f) L. jacobaeus, (g) L. emeroides and (h) L. sessilifolius ........................................................................................................................ 221 Figure 7.2 Liquid chromatography-mass spectrometry (LC-MS) of bee-pollinated flowers (a) Lotus spartioides, a species with yellow flowers that do not modify flower color after anthesis, (b) pre-change yellow flowers of L. sessilifolius, (c) post-change red flowers of L. sessilifolius and (d) a bird-pollinated species, L. berthelotii, with red flowers ....................................... 222 Figure 7.3 Reflectance of yellow bee-pollinated flowers that do not change flower colour (a) Lotus callis-viridis with a UV peak (u+ b- g+ r+), species that modify flower color after anthesis (b) L. latifolius (yellow and red flowers) , (c) L. sessilifolius (yellow and red flowers), and (d) reflectance of L. berthelotii, a bird-pollinated species with red flowers .................. 223 Figure 7.4 Gene expression comparison of three structural genes of the anthocyanin pathway, dihydroflavonol-4-reductase (LjDFR1, 2, 3 and 5), anthocyanin synthase (LjANS) and Omethyl transferase (LjOMT) at mature stages of flower development. Ubiquitin was used as an internal control. Bee-pollinated species of L. japonicus, L. filicaulis and L. sessilifolius, the two latter with pre-change yellow flowers and post-change red flowers. Bird-pollinated species from the rhyncholotus group with red-orange flowers that do not change color after anthesis .............................................................................................................................. 224  xv  Figure 7.5 Schematic representation of the anthocyanin pathway and the major modifications during the evolutionary transition from yellow flowers (bee pollination) to red flowers (bird pollination) in Lotus. Colors at the end of the major pigments indicate the colors produced for each pigment. Bold arrows indicate the pathways active in the bird-pollinated species. The pelargonidin branch is inactive in the three types of petals in this group. Major transitions in pigment composition of bird-pollinated species (1) up-regulation of the cyanidin branch, (2) down regulation of the delphinidin branch and (3) down regulation of flavonol production with a sub sequent modification of flavonol composition. Gray square indicates the branch of the anthocyanin pathway with main pigment production in bird-pollinated species .................. 225 Figure 7.6 Molecular tree based on one nuclear (ITS) and three plastid regions (CYB6, trnH-psbA and matK). Character mapping of the trait flower color change after anthesis in Lotus sections Pedrosia and Rhyncholotus. Red branches show clades where this trait has evolved and the numbers on the tree the times this trait evolved within this group (1-3). Arrows indicate the numbers of reversals, one of which occurred in three species of the rhyncholotus group. (A) represents flower color at anthesis (pre-change) and (B) indicates flower color after change (post-change) ..................................................................................................................... 227  xvi  ACKNOWLEDGEMENTS  I would like to thank my PhD supervisor Quentin Cronk for the guidance and the opportunity to work in his lab during the last five years and a half. I appreciate his mentorship in research, writing and the publication process. I am also grateful for his financial support during my field collections trips during this period. I also would like to thank the members of my committee, Keith Adams, Jeannette Whitton and Wayne Maddison, for his guidance and support during my research. I also thank members of the examination committee for their helpful comments: Kermit Ritland, Sean Graham and Spencer Barret. I am thankful with Dorothy Cheung, Hardeep Rai, Chun-Neng (Bruce) Wang, Nyssa Temmel, Julia Nowak, Christine Wollacot, and Xinxin Xue for their help and support in the lab. I also appreciate the constructive discussions with Ji Yong Yang about flower evolution. I thank all the support from my collaborators, Javier Francisco Ortega, Arnoldo SantosGuerra, Ruth Jaen Molina, Juli Caujape, Felicia Oliva-Tejera and Alfredo Valido, during the field collections. This work was possible thanks to the living collections from the UBC Botanical Garden, Fairchild Tropical Botanical Garden, Montgomery Botanical Garden, Jardín Botánico Regional from CICY, Jardín de Aclimatación de la Orotova and Jardín Canario Botánico Viera y Clavijo. I appreciate the support and the use of these installations during my collection trips. I appreciate the facilities of the CICY, UBC, JAO, and the University of Reading herbaria for allowing the use of their plant material. Seeds and plant material were kindly provided by Brigitte Marazzi, University of Miyazaki and USDA, ARS National Genetic Resources Program. Special thanks to the Cabildos of the Canary Islands (Tenerife, Gran Canaria, El Hierro, La Gomera and La Palma) for providing permits and support during the collection trip. My research work was supported by NSERC Discovery Grant program (Canada) to Quentin Cronk, a PhD scholarship from Consejo Nacional de Ciencia y Tecnología (CONACyT) from Mexico, a Molecular and Organismal Research in Plant History (MORPH) grant, and a graduate fellowship (UGF) from University of British Columbia (UBC). Agradezco profundamente todo el apoyo e inspiracion de SANDRA CERVANTES, el amor de mi vida. No estaria donde estoy si no fuera por ti. Agradezco especialmente el soporte de las familias Arango y Ojeda. xvii  DEDICATION  Dedicated with all my heart to you,  SANDRA and EMILIO  xviii  “Regarding the small size of these islands the sheer amount of endemic species is really remarkable. Furthermore, every mountain is crowned by a young crater and the borders of each lava flow are still clearly recognisable. We have to conclude that not long ago, the ocean was reigning out here. It seems to me, that here in space as well as in time, the secret of all secrets, that is the appearance of new creatures on earth is readily perceptible.”  CHARLES DARWIN, 1845  xix  CO-AUTHORSHIP STATEMENT  Chapter 2 has been published: Cronk, Q. and Ojeda, I. (2008) Bird-pollinated flowers in an evolutionary and molecular context. Journal of Experimental Botany 59: 715-727. The opportunity of this work came as an invitation to Q.C.B. Cronk. The idea came after discussions with Q.C.B. Cronk. The manuscript was written in equal proportion by both authors.  Chapter 3 has been published: Ojeda I, Francisco-Ortega, J. and Cronk, Q. (2009) Evolution of petal epidermal micromorphology in Leguminosae and its use as a marker of petal identity. Annals of Botany 104: 1099-1110. The project was suggested by Q.C.B. Cronk, and I conducted the sampling, data analysis and wrote the manuscript with Q.C.B. Cronk. Francisco-Ortega, J. collaborated in the collection and the analysis of samples from the Fairchild Botanical Garden. He reviewed the draft of the manuscript.  Chapter 4 is a draft of a manuscript that will be submitted for publication: Ojeda, I, SantosGuerra, A., Oliva-Tejera, F., Jaen Molina, R., Francisco-Ortega, J. and Q.C.B. Cronk. The origin of bird pollination in Macaronesian Lotus section Rhyncholotus (Loteae, Leguminosae). The idea originated from discussions with Q.C.B. Cronk. I conducted the sampling, all laboratory work, data analysis and wrote the manuscript. Q.C.B. Cronk provided insights into data analysis and contributed to the manuscript. Santos-Guerra, A., Oliva-Tejera, F., Jaen Molina, R., FranciscoOrtega, J helped during the sampling and field work in the Canary Islands and provided with helpful discussion about the distribution, taxonomy and morphology of the group.  Chapter 5 is a draft of a manuscript that will be submitted for publication: Ojeda, I. SantosGuerra, A., Oliva-Tejera, F. Jaen Molina, R. Caujape, J. Marrero, A. and Cronk. Q. C. B. Barcoding plant island radiations and its applicability in species recognition and conservation in Macaronesian Lotus section Pedrosia and Rhyncholotus (Loteae, Leguminosae). The idea originated from discussions with Q.C.B. Cronk and Hardeep Rai. I conducted the sampling, all laboratory work and data analysis, and wrote the manuscript. Santos-Guerra, A., Oliva-Tejera, F., Jaen Molina, R. Caujape, J. Marrero, A. helped during the sampling and field work in the Canary Islands. They also provided DNA samples from the Jardín Botánico Canario Viera y Clavijo.  xx  Chapter 6 is a draft of a manuscript that will be submitted for publication: Ojeda, I. SantosGuerra, A., Oliva-Tejera, F. Jaen Molina, R. Caujape, J. Marrero, A. and Cronk. Q. C. B. Temporal, but not spatial, changes in expression patterns of floral identity genes are associated with the evolution of bird pollination in Lotus (Leguminosae). The idea originated from discussion with Q.C.B. Cronk. I conducted the sampling, all laboratory work and data analysis, and wrote the manuscript. Santos-Guerra, A., Oliva-Tejera, F., Jaen Molina, R. Caujape, J. Marrero, A. helped during the sampling and field work in the Canary Islands. Jaen-Molina R. helped with RNA extraction in the installations of the molecular biology lab at Jardín Botánico Canario Viera y Clavijo.  Chapter 7 is a draft of a manuscript that will be submitted for publication: Ojeda, I. SantosGuerra, A., Oliva-Tejera, Valido, A., F. Jaen Molina, R. Caujape, Xue, X., J. Marrero, A. and Cronk. Q. C. B. Biochemical evolution of bird-pollinated flowers in Macaronesian Lotus (Leguminosae). The idea and the design of the project originated from discussions with Q.C.B. Cronk. I conducted the sampling, all laboratory work and data analysis, and wrote the manuscript. Santos-Guerra, A., Oliva-Tejera, F., Jaen Molina, R. Caujape, J. Marrero, A. helped during the sampling and field work in the Canary Islands. Jaen-Molina R. helped with RNA extraction in the installations of the molecular biology lab at Jardín Botánico Canario Viera y Clavijo. Xinxin Xue collaborated with pigments extraction and analyses in the lab. Alfredo Valido conducted with me all flower reflectance measurements in the field. All authors contributed reviewing the manuscript.  xxi  1. Introduction 1. 1 Bird pollination Bird pollination or ornithophily has evolved in several plants groups in various regions in the world, especially in the tropics. This coevolutionary adaptation involves the association of several distinctive traits which form the so called ―pollination syndromes‖ (Proctor et al., 1996). The genetic mechanisms by which these floral adaptations have been modified during pollinator shifts have been addressed in a few plant groups. However, the majority of these plant groups have evolved under the selective pressure of hummingbirds in North America (e.g., Mimulus, Aquilegia, Penstemon) (Grant and Grant, 1968). North American bird-pollinated flowers represent a small proportion of the great diversity in floral morphology adapted to birds. Additionally, these bird-pollinated flowers enter into contact with hummingbirds relatively recently, and therefore these model systems do not represent the range of adaptations displayed in ornithophilous plants. This lack of additional groups, in part, motivated the research described in this thesis. In this thesis, I studied how the flower features adapted to bird pollination (opportunistic passerine nectar-feeders) evolved in a group of four species of Lotus (Leguminosae) in the Canary Islands (Macaronesia). I start my thesis (Chapter 2) with a review of what it is currently known about this pollination syndrome. In this chapter, I give a general overview of the bird pollination syndrome. I also summarize the major traits that have been associated with this syndrome (for instance red flowers, nectar composition, and flower shape), with the current knowledge that is known about the candidate genes that control some of these traits. I also discuss some of the models systems used to dissect the genetic factors responsible for these floral traits. I should clarify that there are other plant models not mentioned in this chapter, notably are the cases of Aquilegia (Fang et al., 2010; Kramer, 2009; Kramer and Hodges, 2010) and Petunia (Gübitz et al., 2009). All these 1  plant models will provide in the near future fruitful insights about pollinator transitions and the genetic basis of some of these floral traits.  1.2 Macaronesian Lotus In this thesis, I propose that this group can provide some insight about transitions from be to bird pollination. This group has four species (the ―rhyncholotus group‖) adapted to opportunistic passerine nectar-feeders in the Canary Islands. Although there is no evidence to support the efficiency of opportunistic passerine nectar feeder birds in this plant group, there are additional data, such as nectar composition (Dupont et al., 2004), visitation observation, and other floral traits, that strongly suggest that this group is bird-pollinated (Olesen, 1985; Ollerton et al., 2009; Sletzer, 2005; Valido et al., 2004). The group has a contrasting floral morphology in comparison with its closely relatives. Additionally, the group seems to have evolved recently and crosses between species with contrasting syndrome are likely possible. Therefore, this group represents another model system to address questions in pollination shifts, from a region not represented in previous plant models. Finally, there are a number of genomic and genetic resources from four model systems in legumes (Pisum sativum, Glycine max, Medicago trunculata and Lotus japonicus) which are useful in studying the genetic basis of some of these traits. Of particular interest is L. japonicus which has extensive EST libraries and a whole genome sequence (Sato et al., 2008), tools that are easily transferred to Macaronesian Lotus. In particular, there has been notable advance in the understanding of the genes that determine petal identity in legumes and its association with particular petal micromorphology (Feng et al., 2006b). However, outside the two models species where these features have been  2  characterized (L. japonicus and Pisum sativum) (Feng et al., 2006b; Wang et al., 2008); there was a general lack of the epidermal types in Leguminosae before this thesis. In Chapter 3, I provide a survey of the different epidermal types in a representative group of legumes. This analysis provides a general overview of the major epidermal types within legumes and how some of these traits have evolved within the family. Especially important is the finding that the highly specialized papilionoid flower, with three petal types, is also highly specialized at the micromorphological level, where specific petal surfaces are associated with a particular petal type. This analysis also indicates that this specialization has been lost in some groups in the papilionoids.  1. 3 Molecular phylogenetics and DNA barcoding of Macaronesian Lotus Macaronesian Lotus comprises a group of about 40 species distributed in five volcanic archipelagos and mainland Africa and Europe. The group is high in species diversity and endemics in Morocco and the Canary Islands. Despite the large species numbers within this group, there are relative few phylogenetic analyses within the group and there is a lack of resolution due to low levels of variability of the ITS region and a lack of complete representation of the species. In Chapter 4, I provide a phylogenetic framework of the group where I included the most comprehensive sampling within this group. I used a combined analyses of six gene regions (plastid and nuclear) in three different data sets. I identified one species, Lotus sessilifolius, as the closest relative of the four bird-pollinated species. I also determined that the group evolved relatively recently (within the last 2 Ma) and this radiation is associated with recent volcanic activity in Tenerife and the emergence of La Palma. My analyses suggest that this syndrome is  3  derived in Macaronesian Lotus and may have evolved under the selective pressure of these passerine nectar-feeders (―de novo opportunistic hypothesis‖) (Valido et al., 2004). In Chapter 5, I addressed whether six gene regions (ITS, matK, rpoB, rbcL, rpoC1 and trnH-psbA) have a practical application as barcodes in this group. This legume group represents a particular challenge for barcodes, as some of these clades diverged relatively recently. My analyses suggest that these barcodes have low levels of resolution in comparison with other plants groups that have been analyzed at a floristic level (González et al., 2009; Kress et al., 2009). I was able to identify only 18% at species when I used the recommended barcode combination (rbcL+matK) (CBOL, 2009). When all six regions were combined I identified 52% of the species included and only four (out 10 species) considered endangered within this group.  1.4 Flower colour and petal micromorphology modifications during the transition from bee to bird pollination in Macaronesian Lotus In this thesis, I particularly studied the modification of two floral traits: petal micromorphology and flower colour. In Chapter 6, I addressed the modifications in petal micromorphology between the two pollination syndromes. The transition from bee to bird pollination in this group required the modification of the three types of petals (dorsal, lateral and ventral) within the flower. During the transition the role of each petal is modified within the flower and this modification occurred at the micromorphological level as well. Bee-pollinated species have papillose conical cells on the dorsal and lateral petals, which enhance petal perception and aid grip when pollinators land on the lateral petals. In contrast, all four bird-pollinated species completely lack papillose conical cells and this epidermal type is restricted to only a highly localized area of the lateral petal. This 4  shift in epidermal surface suggests that papillose conical cells may have been lost as an anti-bee strategy and it is only maintained where it enhances bird attraction. I found an association between the micromorphological modifications with the expression patterns of two petal identity genes. These analyses suggest that the micromorphological modifications in the petals may have evolved due to heterochronic changes of expression patterns of two petal identity genes (LjCYC2 and LjCYC3) between the two pollination syndromes. Bird-pollinated flowers have red/orange flowers while the majority of the bee-pollinated species have yellow flowers. In Chapter 7, I explain how red flowers might have evolved in this group. I found that 58% of the bee-pollinated species have the capacity to modify flower color after anthesis. This ability may have evolved in this group and other plant groups as a strategy to increase foraging efficiency (Jones and Cruzan, 1999). I found that this ability has evolved at least three times in Macaronesian Lotus and the rhyncholotus group evolved within a clade where all species have the ability to modify from yellow to red flowers. I found that pre-change yellow flowers differ in spectral reflectance to the post-change red flowers and insect pollinators can distinguish between the two colours. I also show that the pigments found in bird-pollinated flowers are already present in the most closely related species, and the transition required only the re-direction of delphinidin and flavonols towards the cyanidin branch of the anthocyanidin pathway. I suggest that this transition evolved as a heterochronic transition of pigment production and as an anti-bee, rather than a pro-bird strategy.  5  1.5 Bibliography CBOL, 2009. A DNA barcode for land plants. Proceedings of the National Academy of Sciences of the USA 106, 12794-12797. Dupont, Y.L., Hansen, D.M., Rasmussen, J.T., Olesen, J.M., 2004. Evolutionary changes in nectar sugar composition associated with switches between bird and insect pollination: the Canarian bird-flower element revisited. Functional Ecology 18, 670-676. Fang, G.C., Blackmon, B.P., Henry, D.C., Staton, M.E., Saski, C.A., Hodges, S.A., Tomkins, J.P., Luo, H., 2010. Genomic tools development for Aquilegia: construction of a BACbased physical map. BMC Genomics 11, 621. Feng, X., Zhao, Z., Tian, Z., Xu, S., Luo, Y., Ca, Z., Wang, Y., Yang, J., Wang, Z., Weng, L., Chen, J., Zheng, J., Guo, X., Luo, J., Sato, S., Tabata, S., Ma, W., Cao, X., Hu, X., Sun, C., Luo, D., 2006. Control of petal shape and flower zygomorphy in Lotus japonicus. Proceedings of the National Academy of Sciences of the United States of America 103, 4970-4975. González, M.A., Baraloto, C., Engel, J., Mori, S.A., Petronelli, P., 2009. Identification of Amazonian Trees with DNA Barcodes. PLoS ONE 4, e7483. Grant, K.A., Grant, V., 1968. Hummingbirds and their flowers. Columbia University Press, New York and London. Gübitz, T., Hoballah, M.A., Dell’Olivo, A., Kuhlemeier, C., 2009. Petunia as a model system for the genetics and evolution of pollination syndromes. In: Gerarts, T., Strommer, J. (Eds.), Petunia, evolutionary, developmental and physiological genetics. Springer New York, New York, pp. 29-49. Jones, C.E., Cruzan, M.B., 1999. Floral morphological changes and reproductive success in deer weed (Lotus scopiarus, Fabaceae). American Journal of Botany 86, 273–277. 6  Kramer, E.M., 2009. Aquilegia: a new model for plant development, ecology, and evolution. Annual Review of Plant Biology 60, 261-277. Kramer, E.M., Hodges, S.A., 2010. Aquilegia as a model system for the evolution and ecology of petals. Philosophical Transactions of the Royal Society of London B. Biological Sciences 365, 477-490. Kress, W.J., Erickson, D.L., Jones, F.A., Swensond, N.J., Perez, P., Sanjurb, O., Berminghamb, E., 2009. Plant DNA barcodes and a community phylogeny of a tropical forest dynamics plot in Panama. Proceedings of the National Academy of Sciences USA 106, 18621-18626. Olesen, J.M., 1985. The Macaronesian bird flower element and its relation to bird and bee opportunistic. Botanical Journal of the Linnean Society 91, 395-414. Ollerton, J., Cranmer, L., Stelzer, R.J., Sullivan, S., Chittka, L., 2009. Bird pollination of Canary Island endemic plants. Naturwissenschaften 96, 221-232. Proctor, M., Yeo, P., Lack, A., 1996. The natural history of pollination. Timber Press, Portland. Sandral, G., Remizova, M.V., Sokoloff, D.D., 2006. A taxonomic survey of Lotus section Pedrosia (Leguminose, Loteae). Wulfenia 13, 97-192. Sato, S., Nakamura, Y., Kaneko, T., Asamizu, E., Kato, T., Nakao, M., Sasamoto, S., Watanabe, A., Ono, A., Kawashima, K., Fujishiro, T., Katoh, M., Kohara, M., Kishida, Y., Minami, C., Nakayama, S., Nakazaki, N., Shimizu, Y., Shinpo, S., Takahashi, C., Wada, T., Yamada, M., Ohmido, N., Hayashi, M., Fukui, K., Baba, T., Nakamichi, T., Mori, H., Tabata, S., 2008. Genome structure of the legume, Lotus japonicus. DNA Resources 15, 227-239. Sletzer, R., 2005. Sammelstrategien bei Hummeln ein Vergleich zwischen Insel- und Festlandpopulationen Fakultät für Biologie. Universität Würzburg, Würzburg, p. 46. 7  Valido, A., Dupont, Y.L., Olesen, J.M., 2004. Bird-flower interactions in the Macaronesian islands. Journal of Biogeography 31, 1945-1953. Wang, Z., Luo, Y., Li, X., Wang, L., Xu, S., Yang, J., Weng, L., Sato, S., Tabata, S., Ambrose, M., Rameau, C., Feng, X., Hu, X., Luo, D., 2008. Genetic control of floral zygomorphy in pea (Pisum sativum L.). Proceedings of the National Academy of Sciences of the United States of America 105, 10414-10419.  8  2. Bird-pollinated flowers in an evolutionary and molecular context  2.1 Pollination syndromes The concept of pollination syndromes, where specific floral traits are associated with particular pollination mechanisms, dates back to the work of the Neapolitan botanist Federico Delpino (1833-1905). The attraction and utilization of a specific group of animals for pollination, for instance, is associated with specific characteristics of flower morphology, colour, nectar, odour, and orientation (Faegri and van der Pijl, 1966; Fenster et al., 2004; Proctor and Yeo, 1973). However, pollination systems are often more complex than floral morphology would at first sight suggest, and this has led to criticisms of the pollination syndrome concept, mainly based on evidence that flowers attract a broader spectrum of visitors than expected (Waser et al., 1996). Nevertheless, there is ample evidence supporting a strong association between certain floral traits and functional groups of pollinators that exert similar selective pressures (Fenster et al., 2004).  One well-recognized syndrome of floral traits is that associated with bird pollination (ornithophily). Ornithophilous flowers (Fig. 2.1) are very often red with copious dilute nectar. Furthermore they lack characters associated with other pollination syndromes, such as scent.  1  A version of this chapter has been published. Cronk, Q. and Ojeda, I. (2008) Bird-pollinated  flowers in an evolutionary and molecular context. Journal of Experimental Botany 59: 715-727.  9  2.2 The syndrome of ornithophily 2.2.1 The birds Many birds will casually visit flowers in search of food, often primarily to seek insects concealed in inflorescences although they will take nectar if it is available. Flower visiting of some sort has been reported in as many as fifty families of birds (Proctor and Yeo, 1973; Proctor et al., 1996). However, three families of birds have evolved as major groups of flower specialists. These are the hummingbirds (Trochilidae), the sunbirds (Nectariniidae) and the honey-eaters (Meliphagidae) (Fig. 2.2). Hummingbirds are exclusive to the New World ranging from Southern South America to Alaska with the highest diversity in the northern Andes (Grant and Grant, 1968). As hummingbirds visit flowers by hovering there is no need for a perch and in consequence some hummingbird flowers are long and pendulous. Hummingbirds have beaks that are highly specialized for nectar feeding, even though insects form a normal part of their diet, necessitating a remarkable flexibility of the jaw (Yanega and Rubega, 2004). They have undergone a major evolutionary radiation in South America and a secondary radiation in North America (Mayr, 1964). Fossil evidence from Europe, however, suggests that the early evolution of this group was not exclusive to the New World (Mayr, 2004). Sunbirds and spiderhunters (Nectariniidae) are the major group of pollinating birds in Africa and Asia. The honeyeaters (Meliphagidae) are very important pollinators of styphelioid Ericaceae (Epacridaceae), Myrtaceae and Proteaceae in Australia and they extend north to Wallace's line and east to New Zealand and Hawaii. Other important groups include the American orioles (Icteridae) in North and South America, the honeycreepers (Thraupidae) in tropical America and the Hawaiian honeycreepers (Fringillidae, subfamily Drepanidinae) in the Hawaiian Islands. In Africa the White-eyes 10  (Zosteropidae) are another important group, as are the South African sugar-birds (Promeropidae). Bird pollination is particularly common in relatively aseasonal tropical and subtropical regions as flowers and nectar are available year-round to support nectarivorous birds. It tends to be absent or rare in regions in which vegetation has a long dormant period. North America is an exception as here hummingbirds migrate north during the summer. The migration is particularly remarkable on the NW coast where hummingbirds migrate as far as Alaska. Bird pollination is almost entirely absent in Europe and in Asia north of the Himalayas. In Europe, although there are reports that some passeriform birds occasionally feed on nectar (Kay, 1985; Merino and Nogueras, 2003; Proctor et al., 1996; Schwilch et al., 2001), there is only one report of a birdpollinated native plant, Anagyris foetida L. (Leguminosae). In Spain this is apparently pollinated by three warblers, Phylloscopus collybita Vieillot, Sylvia atricapilla L. and S. melanocephala Gmelin (Ortega-Olivencia et al., 2005).  2.2.2 Ecology of bird pollination Some attributes of birds, such as long flight distances and high visual acuity, make them excellent pollinators, especially valuable during inclement weather conditions when other pollinators, such as bees, are inactive. Birds may therefore be important supplemental pollinators in environments where insects have low population densities, such as high altitudes ecosystems (Van der Pijl and Dodson, 1966), dry environments (Stiles, 1978), isolated islands where insect colonization has been poor (Dupont et al., 2004; Micheneau et al., 2006) and for plants flowering during winter months when insects are few (Kunitake et al., 2004). However, birds are large and require more energy than insects. For this reason plants with bird-pollinated flowers tend to put more energy into nectar production and often produce larger 11  flowers to accommodate their avian pollinators. Bird-pollinated plants may also deploy more resources in floral structures that protect against thieves of their abundant nectar (Stiles, 1978). Environments with low plant productivity may be limiting for nectar production and for this syndrome in general. The tropical forest understorey, with limited photosynthetic rates, has relatively few bird-pollinated plants (Stiles, 1978) and the same appears to hold true for cold, hyper-arid and nutrient poor environments. In constrast, the syndrome is particularly common in tropical and subtropical shrublands, open woodland and riverine communities.  2.2.3 Perching vs hovering The behaviour of those birds associated with ornithophily can be broadly divided into two types, hovering and perching. The hovering behaviour is found mainly in hummingbirds, but is also present in some families of passeriforms. In hovering, birds collect nectar without landing on the plant, which may therefore have hanging or pendant flowers. On the other hand, perching birds land on stems, leaf stalks, adjacent branches, flower buds, which must provide an adequate perch. Low herbaceous plants may be pollinated by birds that perch on the ground, and they usually orient their flowers vertically erect. Examples include Lotus berthelotii Masf. and its relatives in the Canary Islands (Olesen, 1985), and Gastrolobium praemorsum (Meisn.) G. Chandler & Crisp in the southwest of Western Australia (Keighery, 1982). Among birds, flower foraging by perching is more widespread and involves fewer specialist adaptations than foraging while hovering. Flower visiting has evolved in several families of perching passeriforms in the New World and the Old World. Unlike hummingbirds, passeriforms tend to forage and travel in groups and can be effective in cross-pollinating even large trees (Stiles, 1981). In the New World, pollination by perching passeriforms appears to have evolved recently, usually, although not always, involving plant species in genera already 12  adapted to hummingbird pollination (Cruden and Toledo, 1977; Toledo, 1975).  2.2.4 Hermit vs non-hermit hummingbirds Bird behaviour is important in determining the nature of bird-plant interaction. The vast majority of bird-pollinated species in the Neotropics, are herbs, shrubs, small trees and epiphytes, and rarely large canopy trees. Solitary behaviour and territoriality in hummingbirds do not allow the levels of pollinator saturation or cross-pollination necessary for the effective pollination of large canopy trees (Stiles, 1975), which therefore tend to be bee-pollinated. Hummingbirds may be divided into two subgroups, hermits and non-hermits, and this division has important implications for pollination. The majority of hermits have long, curved bills and a tendency to forage on flowers with long, curved corollas. They inhabit the understorey of the tropical forest and decrease in abundance and diversity at higher elevations and in dry habitats. Hermits are non-territorial with a traplining method of foraging. Traplining involves visiting many plants sequentially for short visits, flying from plant to plant, often over some distance, and is highly effective in promoting cross pollination (Snow and Snow, 1972; Stiles, 1975, 1981). Non-hermits have short straight bills and a tendency to hold territories and thus nonhermit pollination behaviour favours self-pollination. Non-hermits are widely distributed but have their greatest diversity in the tropical highlands (Snow and Snow, 1972; Stiles, 1975, 1981). The North American hummingbird fauna comprises a fairly homogenous assemblage of nonhermit hummingbirds and probably for this reason North American ornithophilous plants are fairly uniform in floral form compared to those of South America (Grant, 1966; Stiles, 1981). Interestingly, the existence of these two types of hummingbird has had a strong influence on patterns of floral diversity in the genus Heliconia (Heliconiaceae) (Stiles, 1975, 1981). Heliconia 13  species that are pollinated by hermit hummingbirds usually have long, curved corollas, and grow in the forest understorey where light is limiting. They grow in small, scattered clumps and produce a few flowers over a long flowering period. Daily nectar production is low with usually moderate to high nectar concentration (Stiles, 1975), and hermit hummingbirds trapline between clumps. By contrast, non-hermit associated Heliconia species grow in large clumps with highly synchronous flowering, producing many flowers during a definite flowering period. These Heliconia species usually have short straight corollas and are usually found in highly productive environments of forest gaps or along rivers with abundant light (Stiles, 1975). They produce a large quantity of relatively dilute nectar. Territorial non-hermit hummingbirds will defend these clumps because of their abundant nectar reward, which is beneficial for pollination intensity but not for cross-pollination between clumps.  2.2.5 Major groups of angiosperms with bird pollination Bird pollination is widespread in the flowering plants and appears to have evolved many times. It is present in some 65 flowering plant families and in most of these it probably represents a separate origin, usually from a bee-pollinated precursor. However, bird pollination is notably absent in some of the largest families of flowering plants. In Asteraceae, for example, only the South American genus Mutisia (Buzato et al., 2000; Proctor et al., 1996) and Dendroseris litoralis from Juan Fernandez Islands (Bernardello et al., 2001) are bird pollinated. On the other hand, there are some large clades in which bird pollination is particularly common, such as in the monocot order Zingiberales. Families of this order are Cannaceae, Costaceae, Heliconiaceae, Lowiaceae, Marantaceae, Musaceae, Strelitziaceae, Zingiberaceae, and bird pollination occurs in all of them. Genera such as Canna (Cannaceae), Strelitzia (Strelitziaceae), Heliconia (Heliconiaceae) and Costus (Costaceae) are well known for their showy bird14  pollinated species. In the Costaceae, for instance, ornithophily associated with hummingbirds evolved several times in the neotropics from bee-pollinated ancestral species (Specht, 2006). Ornithophily has also evolved several times, mainly from a bee-pollinated ancestor, in many families of the eudicots. A common pattern within these groups is the parallel evolution of the traits associated with ornithophily among non-closely related groups of plants. For example, within Gesneriaceae many Columnea species have similar morphological traits (trailing epiphytes with fairly large and showy red flowers) to another group from the Old World, Aeschynanthus. Columnea is distributed in Central and South America and is pollinated by hummingbirds, whereas Aeschynanthus is distributed in the Palaeotropics. The multiple origin of bird pollination in the flowering plants raises the question of what pre-adaptations promote this evolutionary transition. Two features, commonly associated with bee pollination, appear to be permissive of a transition to bird pollination. One is floral zygomorphy (monosymmetry or asymmetry) and the other is the possession of a floral tube. Bird pollination is common in many families with strongly zygomorphic flowers (e.g. Scrophulariaceae, Heliconiaceae, Gesneriaceae and Leguminosae). It is generally considered to be an adaptation to bee pollination and allows precise placement of pollen on the body of the bee. It seems that these adaptations may be important for bird-pollination also. The same applies to the character of perianth segments being connate into a tubular corolla (or the presence of a tubular hypanthium in Passiflora and Fuchsia). Tubular flower originally evolved for insect pollination but are equally suitable for the probing foraging of bird’s beak. Hummingbird flowers in North America are predominantly sympetalous dicots and it has been suggested that this corolla condition, with its tubular shape, is another pre-adaptation to hummingbird pollination (Grant and Grant, 1968).  15  2.3 The floral phenotype 2.3.1 Types of floral adaptation Floral adaptations to bird pollination fall into four broad types (1) attraction mechanisms, (2) exclusion mechanisms, (3) protection mechanisms, and (4) pollination mechanisms (Grant and Grant, 1968). Attraction mechanisms are those such as copious nectar and vivid floral display that attract birds to flowers. The floral display may be just red or orange, or a combination of contrasting colours, including orange, yellow, green and blue ("parrot colours"). Strelitzia reginae W. Aiton, for instance, presents a striking display of orange and blue. Exclusion mechanisms are those features that help to deter illegitimate flower visitors that might otherwise interfere with pollination and rob nectar. Red colour, long and narrow floral tubes and the absence of insect landing platforms are the most obvious of these and are further discussed below. Pendent flowers, as in Aquilegia formosa Fisch. ex DC., are difficult for insects to work but easy for hummingbirds. Recurving petals, as in Ipomopsis aggregata (Pursh) V. Grant, or the short and recurved lower lip of Mimulus cardinalis Dougl. ex Benth., serve to deny insects a landing platform and similarly make the flowers difficult to work except by hummingbirds. In Trichostema lanatum Benth., the long-exserted stamens are recurved to block entrance to the tube. Protection mechanisms are also important, as birds are large and potentially destructive pollinators. A common form of protection is provided by mechanical strengthening of the flower by the formation of sclerenchyma or collenchyma tissue in various floral parts. The ovary and ovules, often situated near to the nectaries are particularly vulnerable to the probing of bird's beaks. Protection of ovules takes many forms. There may be separation of ovary and nectary, either by the sheathing of the ovary by a staminal tube, or by a stalked or inferior ovary. Alternatively there may be a groove formed by the corolla to guide birds' beaks to the nectary 16  without causing damage, or ridges of the corolla to provide direct protection to the ovary. In addition, the style may be protected in a groove formed from rides of the upper petals, as in Justicia californica (Benth.) D. Gibson. In Penstemon it is the nectaries rather than the ovary that shift. In bird-pollinated species of this genus, nectaries are displaced upwards from the base of the ovary to the outer bases of the upper pair of stamens. Pollination mechanisms are those that enhance the precise deposition of pollen on bird and stigma. These include both spatial and temporal relations of the reproductive organs to the position of pollinating birds. The long exserted stamens of bird-pollinated species of Aquilegia (e.g. A. formosa), Fuchsia and Ribes (e.g. R. speciosum Pursh) dust birds' heads or even backs, with pollen. In the radially symmetrical Ipomopsis aggregata the ring of five stamens place pollen all around the base of the birds' beaks. Zygomorphic flowers, on the other hand, tend to place pollen on the top of the beak or on the top of the bird's head.  2.3.2 Why red? Explanations for the remarkably consistent association of bird pollination with red or reddish flowers take two forms, either avoidance of bees (and other insect pollinators), or attraction of birds. It seems that red colour is not necessary to attract birds. There are examples known where birds are effective pollinators of species with orange, yellow and white flowers (and less frequently reddish violet and blue flowers) (Micheneau et al., 2006; Ortega-Olivencia et al., 2005; Proctor and Yeo, 1973). This has led to the suggestion that avoidance of bees (which cannot see red) is more important than attraction of birds (Proctor and Yeo, 1973; Proctor et al., 1996). Birds perceive colour over wavelengths ranging between 300 and 660 nm, whereas bee vision is in the range 300 to 550 nm. In the neotropical forests bird-pollinated flowers have been shown to have typical median reflectance greater that 585 nm, outside the visual range of bees 17  (Altshuler, 2003). However it should be noted that bees can perceive (and do visit) some flowers seen as red by humans, if they have at least some reflectance in the shorter wavelengths as well (Chittka and Waser, 1997). However, as well as its invisibility to bees, the fact that red is very readily detectable by birds is also likely to be significant. For instance the visual prominence of red may be important for migratory hummingbirds, which can easily detect red flowers when entering a new habitat and associate it with reward (Grant, 1966). Indeed, the association of red flowers and bird pollination may be explained via optimal foraging theory and the relative efficiency of bees and birds in detecting red (Rodriguez-Gironés and Santamaría, 2004), with red flowers acting as a signal of high caloric reward (Raven, 1972). With the association between high energetic rewards and the colour red already well established, there may be strong selective pressure for other bird-pollinated flowers to adopt this "common advertising strategy". However, there is a curious twist in Fuchsia excorticata L.f. of New Zealand, in which a developmental change, from green to red flower colour, signals lack of nectar reward in post-reproductive flowers (Delph and Lively, 1989). Birds therefore avoid the red flowers in favour of green.  2.3.3 Flower colour and its perception Flowers that appear red to humans are of three types (Chittka and Waser, 1997). Some species have an additional peak that stimulates the blue receptor of bees; for example, the red flowers of Dianthus carthusianorum L. are perceived as blue-bee, as this species has a reflectance peak that stimulates the blue receptor of bees. The red flowers of field poppies (Papaver rhoeas L.) have a reflectance peak below 400 nm and are perceived by bees as ultraviolet. However, typical red hummingbird flowers, such as Ipomopsis aggregata and Justicia rizzinii Wassh., will be perceived as green by bees (Chittka et al., 1994; Chittka and 18  Waser, 1997). These flowers only stimulate the green receptor of bees, and there is no additional peak of reflectance in other wavelengths. This type of flower, with a peak only in the red, and with no blue or UV reflection, is more difficult to detect than flowers of other colours (Chittka et al., 1994; Chittka and Waser, 1997). On the other hand, birds are tetrachromatic and they have an additional UV receptor, compared to humans making colour perception in birds more complex than in mammals (Bowmaker, 1977). Further, they have oil droplets that can act as filters that increase the complexity of colour perception. Birds can see in theory twice the number of colours compared with trichromats (Ödeen and Håstad, 2003). It seems that birds do not have an innate preference for red colour, although most of them have their greatest spectral sensitivity and hue discrimination towards the long wavelength end of the spectrum (Stiles, 1981). Experiments with hummingbirds have shown that they learn to associate a range of colours with rewards and that this behaviour can be modified (Proctor et al., 1996). Floral pigments (anthocyanidins and flavonols) have the major impact on the wavelength of light reflected from flowers. There are three major types of anthocyanidin pigments: pelargonidin (generally red), cyanidin (typically magenta or blue depending on pH) and delphinidin (generally blue). Bird-pollinated flowers are much more likely to contain pelargonidin and much less likely to contain delphinidin than flowers generally (Scogin, 1988). A general predominance of pelargonidin in tropical floras (Beale et al., 1941) has been attributed to the tropical distribution of hummingbird pollination (Harborne, 1976). A difference in pigment composition has been reported in comparisons between perching-bird and hummingbird-visited flowers (Scogin, 1988), with a greater prevalence of cyanidin, as opposed to pelargonidin, in perching-bird flowers.  19  2.3.4 Nectar Bird-pollinated flowers generally produce a large quantity of dilute nectar as the main pollinator reward. Nectar characteristics, such as (1) volume, (2) sugar concentration and viscosity, (3) sugar composition, and (4) amino acid composition, are extremely important in determining the success of plant-pollinator interactions (Baker and Baker, 1983a; Proctor et al., 1996) and nectars of bird-pollinated plants tend to be recognizable as such. The volume of nectar in flowers is generally correlated with the size of the flower (Baker, 1978). However, size for size, bird-pollinated plants tend to secrete larger quantities of nectar relative to bee-pollinated species. The same is true in bat-pollinated species. Some bat-pollinated Bombax species, for example, can produce between 200-300 µl of nectar daily (Baker, 1978). Nectar concentration is a characteristic often inversely linked to nectar volume. This holds true in bird-pollinated flowers, which produce relatively dilute nectars but in large quantities. Mean sugar concentrations in nectars of bird-pollinated flowers range between 20 and 26% (Pike and Waser, 1981; Proctor et al., 1996; Stiles and Freeman, 1993) with extremes between 10 and 34% (Baker, 1975). The sugar concentration of nectar determines its viscosity, which is an important physical property that is thought to affect the ease of uptake of nectar by birds. Large quantities of nectar at low concentrations have therefore been explained on the basis of constraints of its uptake associated with viscosity (Baker, 1975), and nectar viscosity tends to remains constant even under different environmental conditions (Proctor et al., 1996). However, dilute nectar with low sugar concentration is less optimal for bees, and so this trait may be more anti-bee than pro-bird (Bolten and Feinsinger, 1978). Nectar with sugar concentration below 18% is not beneficial to honeybees because of the high energetic cost of evaporating water in order to produce honey (Percival, 1965).  20  Nectar sugar composition is also important. Nectar is mainly composed of fruit sugars such as glucose and fructose (hexoses), and/or the disaccharide sucrose. Based on composition, the following types of nectar have been distinguished: (1) sucrose-dominant nectar, (2) sucroserich, (3) hexose-rich, and (4) hexose-dominant (Baker and Baker, 1983a). Hummingbird pollinated flowers generally have sucrose-dominant nectar, whereas flowers pollinated by passerine perching birds tend to have hexose-dominant nectar (Baker and Baker, 1983b, 1990; Baker et al., 1998; Elisens and Freeman, 1988; Freeman et al., 1984; Lammers and Freeman, 1986; Perret et al., 2001; Stiles and Freeman, 1993). Bee-pollinated flowers characteristically have sucrose-rich or sucrose-dominant nectar (Baker and Baker, 1983a), so such nectar composition in hummingbird-pollinated flowers that have evolved from bee-pollinated flowers is not surprising. More difficult to explain are the hexose-rich nectars. Passerine birds often feed on fruits as well, and it has been suggested that the ―taste‖ for hexose fruit sugars derived from this (Baker and Baker, 1983b). Also, there is evidence that some Old World birds (such as Sturnidae, which includes some facultative nectarivores) have difficulty digesting sucrose (Schuler, 1977). Amino acids are another important constituent of nectar. They provide a source of protein-producing substances, perhaps affect the ―taste‖ of the nectar (Baker and Baker, 1983a) or stabilize the sugars in the nectar, i.e. to avoid crystallization (Baker and Baker, 1983a). Surveys have reported between 2 to 24 amino acids in nectar from different species (Baker, 1978). Bird-pollinated flowers contain low concentration of amino acids, which is mainly because birds have additional sources of amino acids (Baker and Baker, 1986). One exception to this trend has been studied in Erythrina (Leguminosae), which is pollinated by territorial orioles and tanagers, whose uptake of amino acids mainly depend on this plant (Baker and Baker, 1982).  21  2.3.5 Corolla morphology in bird-pollinated flowers Flower morphology associated with ornithophily may be divided into five major groups (Proctor et al., 1996). In the brush-flower group, the flowers are arranged in clusters, usually in spheres or cylinders, with protruding stamens, and pollen placement is generalized. The Australian flora is particularly notable for its brush flowers, for instance in Proteaceae and in Acacia species (Lara and Ornelas, 2001). Another common flower trait is the hanging bell, producing copious nectar. Examples of this are found in Canarina (Campanulaceae) (Dupont et al., 2004; Valido et al., 2004), Fritillaria (Liliaceae) (Burquez, 1989; Peters et al., 1995) and Cadia (Fabaceae) (Citerne et al., 2006). However, many ornithophilous flowers have long sympetalous corolla tubes. Such corolla tubes differ from those that are bee pollinated by being long and narrow (the shape of a bird’s beak rather than a bee’s body) and the inconspicuous size and position of the corolla lobes. Rather than forming a landing platform, as in bee-pollinated flowers, the lower corolla lobes are often reflexed under the flower so preventing alighting by insects. This has been termed the "dogfish" flower-type (Proctor and Yeo, 1973) from a fancied resemblance to the backwardssloping snout of elasmobranch fish. The long, narrow tubular corolla is important in deterring bees from accessing nectar. However, this trait may be circumvented by nectar robbing (Lara and Ornelas, 2001) if the base of the corolla tube is not protected by a robust calyx. The corolla tube may also be downcurved (arcuate) in a similar way to many bird beaks. There is also a benefit to being downwardly directed in that insect access is made more difficult. However, most North American hummingbirds have relatively short straight beaks and, probably in consequence, most North American ornithophilous flowers have rather similar short straight and narrow corolla tubes.  22  Differences between corolla tubes of bee and bird-pollinated species are well illustrated in the genus Streptocarpus (Gesneriaceae) (Harrison et al., 1999; Hughes et al., 2007b). Between the bee-pollinated S. rexii Lindl. and the bird-pollinated S. dunnii Mast. there are five major floral differences associated with pollination syndrome. S. dunnii has small flowers that are massed into large inflorescences, while in S. rexii the floral display is provided mainly by large individual flowers. Streptocarpus dunnii has a cylindrical corolla tube suitable for probing by a bird’s beak, whereas S. rexii has an open funnel-shaped flower allowing for ingress by relatively large bees. The orientation of the flowers is varied in S. rexii, while in S. dunnii all the flowers on the inflorescence face the single, large leaf that probably serves as a perching surface for birds. S. dunnii has red flowers without nectar-guides, whereas S. rexii has pale violet flowers marked with the dark nectar-guides that are often found in bee pollination. Finally, S. dunnii has a slightly longer adaxial than abaxial corolla tube length, producing a straight or slightly downcurved flower. This morphology is suitable for an ascending approach by a bird perching on the foliage below. S. rexii, on the other hand, has a shorter adaxial than abaxial corolla length, giving a swept-back flower, so providing an insect landing platform.  2.3.6 Other characters associated with bird pollination. Whereas red flower colour and characteristic nectar are nearly universal in ornithophily, a number of other characters are more minor or associated only with specific examples of ornithophily. These include: (1) floral posture (2) secondary perches (3) protection (4) floral clustering and (5) prominent nectar and (6) absent characters (1) Floral posture is closely correlated to pollinator behaviour. Nodding flowers without perches such as Fuchsia and bird-pollinated Aquilegia species are invariably hummingbird pollinated, as hummingbirds have the ability to hover under a downward facing flower and direct 23  their beaks upward into what is generally a long nectar path. This posture effectively excludes other pollinators by making the flower difficult to access. Bee flowers, by contrast, are often horizontally oriented, in accord with the generally horizontal approach flight of bees, and have horizontal or near-horizontal lower lips for alighting. Flowers pollinated by perching birds that probe from above for nectar are often upwardly oriented. The genus Tillandsia provides examples of both types: short inflorescences with upwardly directed flowers for foraging by birds perching on the stiff leaves, or long arching inflorescences bearing pendulous flowers for hummingbird pollination. The bird-pollinated Streptocarpus dunnii has a single leaf, which probably functions as a perch for foraging birds and the flowers are unidirectionally oriented and down-curved towards the single leaf. Related bee-pollinated Streptocarpus species have flowers oriented in all directions. The South African Phygelius capensis E. Mey. ex Benth. has flowers that are somewhat resupinate, turning back towards the stems from where they may be easily probed by perching birds. In contrast insectpollinated and hummingbird pollinated flowers tend to face outwards towards the incoming flight path of the pollinators. (2) Secondary perches are important for facilitating pollination by birds requiring a perch for floral foraging. The most striking example is provided by the 'rat's tail' babiana (Babiana ringens Ker Gawl.) (Anderson et al., 2005). This produces completely sterile robust inflorescence stalks functioning exclusively to provide a perch for foraging by the malachite sunbird (Nectarinia famosa (L.)). The stiff inflorescence bract of Strelitzia provides a strong perch for birds working the Strelitzia flowers with their feet. (3) Protection. Large vertebrate pollinators can be damaging to all but the most robust of flowers. Perching birds are frequently highly destructive of flowers and may destroy them in the search for insects and nectar. For this reason many flowers have protection in the form of tough 24  construction of parts. Strelitzia reginae is an example of a flower in which the floral parts are rather cartilaginous in texture and robust enough to survive rough foraging by pollinating birds. (4) Floral clustering. Dense inflorescences are frequently associated with pollination by perching birds in the Old World. Rather than flying from single flower to single flower as bees and hummingbirds do, perching foragers will exploit flower clusters by probing several flowers from the same perch. A good example is provided by the robust, multi-flowered inflorescences of Banksia, which are pollinated by honey-eaters and other animals. (5) Prominent nectar is frequently associated with bird pollination in those flowers wide enough for the nectar to be seen. Birds have great visual acuity and often nectar is presented to birds with striking visual cues. The "nectar globes" often found at the base of hanging bell shaped flowers are an example of this. In Fritillaria imperialis L., a plant known since the 18th century to attract birds to feed on the nectar (White, 1789), these take the form of prominent white depressions in the bases of the perianth segments that fill with large drops of nectar. Nesocodon mauritianus (I.B.K. Richardson) Thulin has nectar that is prominent for a different reason, it is coloured red with an aurone pigment, a phenomenon which is surprisingly common and generally associated with bird or reptile pollination (Hansen et al., 2006; Olesen et al., 1998). (6) Characters conspicuous by their absence. These include scent, night blooming and nectar guides. In contrast to bats, which visit flowers that often have a rancid or mousey smell, many birds appear to have a poor sense of smell and bird-pollinated flowers are usually odourless. Not all birds have poor olfaction and exceptions include vultures, tube-nosed Procellaridae (storm petrels and albatrosses) and the nocturnal kiwis (Apteryx spp), all of which rely heavily on olfaction for foraging. However, nectar-feeding birds do not belong to this type and appear to forage by sight, and strictly diurnally. Nectar guides (markings under visible or 25  ultraviolet light that guide insects to floral rewards) are generally absent in bird-pollinated flowers. This is unsurprising as while the insect compound eye is optically crude, bird vision is excellent (with good depth perception) rendering foraging within flowers easy.  2.4 The molecular basis for the evolution of ornithophily The bird-pollination syndrome appears to have evolved independently many times in a wide variety of plant families and genera (Stiles, 1981; Thomson et al., 2000). The syndrome may be initiated with birds experimentally foraging for insects or other sources of food into the flowers (Proctor and Yeo, 1973; Proctor et al., 1996), followed by selection of mutations that increase flower visitation and effectiveness of pollen transfer by birds. The molecular basis for these mutations has been addressed using model systems, in which a shift from one syndrome to another is examined in closely related species with contrasting flower morphology.  2.4.1 Some model systems Ipomoea. The genus Ipomoea (morning glories) has tubular flowers in a wide variety of colours associated with different pollinators. The ancestral colour in this group is blue/purple and together with other traits (broad floral tube, moderate nectar production, inserted stigma and nonversatile anthers) this indicates an adaptation to bee pollination (McDonald, 1991). In one clade, however, there has been a shift to red flowers and hummingbird pollination, encompassing some six species (including I. quamoclit L.). Generally, the anthocyanin cyanidin, and its derivatives, produces purple flowers whereas pelargonidin and its derivatives result in red flowers. The gene flavonoid-3’-hydroxylase (F3’H) is important in controlling pigment production, and the flower colour shift in the I. quamoclit lineage has been shown to be due to a down-regulation of the F3’H enzyme (Zufall and Rausher, 2003, 2004). F3'H is directly responsible for the 26  hydroxylation of anthocyanidin precursors at the 3' position that is required for the production of cyanidin rather than pelargonidin. Furthermore the enzyme dihydroflavonol 4-reductase (DFR), with a role downstream of F3’H, seems to have lost its substrate affinity in this species. Although the major change in this case is a shift in flower colour, other flower traits, such as stigma position, flower tube width and the amount of nectar produced, also distinguish the two pollination syndromes. Mimulus. In this genus two closely related species, M. lewisii Pursh and M. cardinalis Douglas ex Benth., display a great differences in floral characteristics. The former is pollinated mainly by bumblebees and has pink flowers with a wide corolla, nectar guides, and a modest amount of nectar; the latter is associated with hummingbird pollination and has red flowers with reflexed corolla lobes, a narrow corolla tube, and copious production of nectar (Ramsey et al., 2003). A major QTL for flower colour is associated with the allele YELLOW UPPER (YUP) that controls yellow pigment concentration (Bradshaw et al., 1998; Bradshaw et al., 1995). In M. lewisii the dominant allele YUP prevents carotenoid deposition, thus petals have only their pink anthocyanin pigments; whereas in M. cardinalis, the recessive allele yup allows carotenoid deposition and flowers with red colouration. Flower colour is known to have a very marked effect on pollinator visitation (Schemske and Bradshaw, 1999), and when the allele of M. cardinalis yup allele is introgressed into the M. lewisii background, hummingbird visitation increases dramatically, whereas bee visitation is considerably lowered (Bradshaw and Schemske, 2003). This suggests that an adaptive divergence in pollination syndrome can be initiated by a major change in flower colour alone. No gene has yet been cloned as responsible for YUP and its molecular identity still remains unknown. Lotus. Lotus is a large genus in the Fabaceae, containing about 130 species divided into 14 sections. Lotus flowers are zygomorphic and typical of the papilionoid legumes. There are 27  five petals: one dorsal (the standard) usually large and conspicuous, two lateral petals (wings) and two ventral petals (keel petals) that enclose the stamens and the ovary. Pollination takes place when a bee lands and depresses the wings and the keel, forcing out a string of pollen from the stamens located beneath the keel and placing it in the underside of the visitor. Other floral features of Lotus that are associated with bee pollination are the horizontal position, yellow colour, sucrose-dominant nectar composition and scent production. Bird pollination occurs in four species of Lotus from the Canary Islands (Olesen, 1985; Valido et al., 2004). In this group (also known as Rhyncholotus) the flowers are red-orange, in a vertical position, with hexose-dominant nectar and no scent production (Dupont et al., 2004). However, the most striking differences are the size and shape of the petals (Fig. 2.3). In the birdpollinated species the flowers are about twice the size of those of typical bee-pollinated species and the dorsal petal is bent backwards while the ventral petals point up. The floral mechanism is effective in depositing pollen either on the top of the head or on the throat of a foraging bird (Fig. 2. 3). To date, two bird species, the Canarian chiffchaff Phylloscupus canariensis and the Blue tit Parus caeruleus have been reported feeding on some species of this group (Ollerton et al., 2009; Stelzer, 2005). This system is a promising one in which to study the role of petal identity in transitions between pollination systems. The genes responsible for dorsal petal identity have been identified in Lotus japonicus (Regel) K. Larsen. These are the legume CYCLOIDEA- like genes (Citerne et al., 2003; Cronk, 2006b; Feng et al., 2006a), which have already been shown to be involved in the shift to bird-pollination in the papilionoid legume Cadia (Citerne et al., 2006).  28  Figure 2.1 Forms of bird-pollinated flowers. (A) Strelitzia reginae. (B) Erythrina suberosa, (C) Babiana ringens with sterile inflorescences for perching birds (arrow). (D) Cadia purpurea, a member of the genistioids with radial symmetry and nectar globes (arrow). (E) Ipomopsis aggregata. (F) Phygelius capensis. (G) Psittacanthus sp. (H) Fritillaria suberosa with nectar globes (arrow).  A  E  C  B  F  G  D  H  29  Figure 2.2 Approximate world distributions of the three main families of flower visiting birds: hummingbirds (Trochilidae), sunbirds (Nectariniidae), and honey-eaters (Meliphagidae).  30  Figure 2.3 Flowers of Lotus species. Bird-pollinated species of (A) Lotus berthelotii Masf. and (B) L. maculatus Breitfeld, two members of the subgenus Pedrosia s.l. (D) The model legume L. japonicus GIFU B-129 and (E) L. arenarius Brot. a closely related species of the bird-pollinated species within the subgenus Pedrosia s.l. (C, F) Diagrammatic representation of the hypothetical mechanism by which a bird seeks nectar in flowers of L. berthelotii. Arrows indicate the direction in which the dorsal petal is pushed. a, Dorsal petal; b, lateral petal; and c, ventral petal. (C, F) Modified from Olesen (1985).  A  C  B  b c  a a  b  b  c  c  a  E  D  F  a  a c b c  b  a  c b  31  2.6 Bibliography Altshuler, D.L., 2003. Flower color, hummingbird pollination, and habitat irradiance in four Neotropical forests. Biotropica 35, 344-355. Anderson, B., Cole, W.W., Barrett, S.C.H., 2005. Specialized bird perch aids cross-pollination. Nature 435, 41-42. Baker, H., 1975. Sugar concentration in nectars from hummingbird flowers. Biotropica 7, 37-41. Baker, H., 1978. Chemical aspects of the pollination biology of woody plants in the tropics. In: Tomlinson, P., Zimmermann, M. (Eds.), Tropical trees as living systems. Cambridge University Press, Cambridge, pp. 57-82. Baker, H., Baker, I., 1982. Chemical constituents of nectar in relation to pollination mechanism and phylogeny. In: Nitecki, M.H. (Ed.), Biochemical aspects of evolutionary biology. University of Chicago Press, Chicago. Baker, H., Baker, I., 1983a. A brief historical review of the chemistry of floral nectar. In: Bentley, B., Elias, T. (Eds.), The biology of nectaries. Columbia University Press, New York, pp. 126-152. Baker, H., Baker, I., 1983b. Floral nectar sugar constituents in relation to pollination type. In: Jones, C.E., Little, R.J. (Eds.), Handbook of experimental pollination biology. Van Nostrand Reinhold Company Inc, New York. Baker, H., Baker, I., 1986. The occurrence and significance of amino acids in floral nectar. Plant Systematics and Evolution 151, 175-186. Baker, H., Baker, I., 1990. The predictive value of nectar chemistry to the recognition of pollinator types. Israel Journal of Botany 39, 157-166. Baker, H., Baker, I., Hodges, 1998. Sugar composition of nectar and fruits consumed by birds and bats in the tropics and subtropics. Biotropica 30, 559-586. 32  Beale, G.H., Price, J.R., Sturgess, V.C., 1941. A survey of anthocyanidins. VII. The natural selection of flower color. Proceedings of the Royal Society of London. Series B, Biological Sciences 130, 113-126. Bernardello, G., Anderson, G.J., Stuessy, T.F., Crawford, D.J., 2001. A survey of floral traits, breeding systems, floral visitors, and pollination systems of the angiosperms of the Juan Fernández Islands (Chile). The Botanical Review 6, 255-308. Bolten, A.B., Feinsinger, P., 1978. Why do hummingbird flowers secrete dilute nectar? . Biotropica 10, 307-309. Bowmaker, J.K., 1977. The visual pigments, oil droplets, and spectral sensitivity of the pigeon. Vision Research 17, 1129-1138. Bradshaw, H.D.J., Otto, K.G., Frewen, B.E., McKay, J.K., Schemske, D.W., 1998. Quantitative trait loci affecting differences in floral morphology between two species of monkeyflower (Mimulus). Genetics 149, 367-382. Bradshaw, H.D.J., Schemske, D.W., 2003. Allele substitution at a flower color locus produces a pollinator shift in monkeyflowers. Nature 426, 176-178. Bradshaw, H.D.J., Wilbert, S.M., Otto, K.G., Schemske, D.W., 1995. Genetic mapping of floral traits associated with reproductive isolation in monkeyflowers (Mimulus). Nature 376, 762-765. Burquez, A., 1989. Blue tits, Parus caeruleus, as pollinators of the crown imperial, Fritillaria imperialis in Britain. Oikos 55, 335-340. Buzato, S., Sazima, M., Sazima, S., 2000. Hummingbird-pollinated floras at three Atlantic forest sites. Biotropica 32, 824-831. Chittka, L., Shmida, A., Troje, N., Menzel, R., 1994. Ultraviolet as a component of flower reflections, and the color-perception of Hymenoptera. Vision Research 34, 1489-1508. 33  Chittka, L., Waser, N.M., 1997. Why red flowers are not invisible to bees. Israel Journal of Plant Sciences. 45, 169-183. Citerne, H.L., Luo, D., Pennington, R.T., Coen, E., Cronk, Q.C.B., 2003. A phylogenomic investigation of CYCLOIDEA-like TCP genes in the Leguminosae. Plant Physiology 131, 1042-1053. Citerne, H.L., Pennington, R.T., Cronk, Q.C., 2006. An apparent reversal in floral symmetry in the legume Cadia is a homeotic transformation. Proceedings of the National Academy of Sciences United States of America 103, 12017-12020. Cronk, Q.C., 2006. Legume flowers bear fruit. . Proceedings of the National Academy of Sciences United States of America 103, 4801-4802. Cruden, R.W., Toledo, V.M., 1977. Oriole pollination of Erythrina breviflora (Leguminosae): evidence form a polytypic view of ornithophily. Plant Systematics and Evolution 126, 393403. Delph, L.F., Lively, C.M., 1989. The evolution of floral color change: pollinator attraction versus physiological constraints in Fuchsia excorticata. Evolution 43, 1252-1262. Dupont, Y.L., Hansen, D.M., Rasmussen, J.T., Olesen, J.M., 2004. Evolutionary changes in nectar sugar composition associated with switches between bird and insect pollination: the Canarian bird-flower element revisited. Functional Ecology 18, 670-676. Elisens, W.J., Freeman, C.E., 1988. Floral nectar sugar composition and pollinator type among New World genera in tribe Antirrhineae (Scrophulariaceae). American Journal of Botany 75, 971-978. Faegri, K., van der Pijl, L., 1966. The principles of pollination ecology. Pergamon, Oxford. Feng, X., Zhao, Z., Tian, Z., Xu, S., Luo, Y., Ca, Z., Wang, Y., Yang, J., Wang, Z., Weng, L., Chen, J., Zheng, J., Guo, X., Luo, J., Sato, S., Tabata, S., Ma, W., Cao, X., Hu, X., Sun, C., 34  Luo, D., 2006. Control of petal shape and flower zygomorphy in Lotus japonicus. Proceedings of the Natural Academy of Sciences of the United States of America 103, 4970-4975. Fenster, C.B., Armbruster, W.S., P., W., Dudash, M.R., Thompson, J.D., 2004. Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution, and Systematics 35, 375-403. Freeman, C.E., Reid, W.H., Becvar, J.E., Scogin, R., 1984. Similarity and apparent convergence in the nectar-sugar composition of some hummingbird-pollinated flowers. Botanical Gazette 145, 132-135. Grant, K.A., 1966. A hypothesis concerning the prevalence of red coloration in California hummingbird flowers. The American Naturalist 100, 85-97. Grant, K.A., Grant, V., 1968. Hummingbirds and their flowers. Columbia University Press, New York and London. Hansen, D.M., Beer, K., Muller, C.B., 2006. Mauritian coloured nectar no longer a mystery: a visual signal for lizard pollinators. Biology Letters 2, 165-168. Harborne, J.B., 1976. Functions of flavonoids in plants. In: Goodwin, T.W. (Ed.), Chemistry and biochemistry of plants. Academic Press, New York. Harrison, C.J., Möller, M., Cronk, Q.C., 1999. Evolution and Development of Floral Diversity in Streptocarpus and Saintpaulia. Annals of Botany 84, 49-60. Hughes, M., Möller, M., Edwards, T.J., Bellstedt, D.U., W., V., 2007. The impact of pollination syndrome and habitat on gene flow: a comparative study of two Streptocarpus (Gesneriaceae) species. American Journal of Botany 94, 1688-1695. Kay, K.A., 1985. Nectar from willow catkins as a food source of blue tits. Bird study 32, 40-44.  35  Keighery, G.J., 1982. Bird-pollinated plants in Western Australia. In: Amnstrong, J.A., Powell, J.M., Richards, A.J. (Eds.), Pollination and evolution. Royal Botanical Garden, Sidney. Kunitake, Y.K., Hasegawa, M., Myyashita, T., Higuchi, H., 2004. Role of seasonally specialist bird Zosterops japonica on pollen transfer and reproductive success of Camellia japonica in a temperate area. Plant Species Biology 19. Lammers, T.G., Freeman, C.E., 1986. Ornithophily among the Hawaiian Lobelioideae (Campanulaceae): evidence from floral nectar sugar composition. American Journal of Botany 73, 1613-1619. Lara, C., Ornelas, J.F., 2001. Preferential nectar robbing of flowers with long corollas: experimental studies of two hummingbird species visiting three plant species. Oecologia 128, 263-273. Mayr, E., 1964. Inferences concerning the Tertiary American bird faunas Proceedings of the Natural Academy of Sciences USA 51, 80-88. Mayr, G., 2004. Old World fossil record of modern-type hummingbirds. Science 304, 861-864. McDonald, A., 1991. Origin and diversity of Mexican Convolvulaceae. Anales del Instituto de Biología de la UNAM Serie Botánica 62, 65-86. Merino, S., Nogueras, I., 2003. Loquat Eriobotrya japonica as a winter nectar source of birds in central Spain. Ardeola 50, 265-267. Micheneau, C., Fournel, J., Pailler, T., 2006. Bird pollination in an angraecoid orchid on Reunion Island (Mascarene Archipelago, Indian Ocean). Annals of Botany 97, 965-974. Ödeen, A., Håstad, O., 2003. Complex distribution of avian color vision systems revealed by sequencing the SWS1 opsin from total DNA. Molecular Biology and Evolution 20, 855861.  36  Olesen, J.M., 1985. The Macaronesian bird flower element and its relation to bird and bee opportunistic. Botanical Journal of the Linnean Society 91, 395-414. Olesen, J.M., Ronsted, N., Tolderlund, U., Cornett, C., Molgaard, P., Madsen, J., Jones, C.G., Olsen, C.E., 1998. Mauritian red nectar remains a mystery. Nature 393, 529-529. Ollerton, J., Cranmer, L., Stelzer, R.J., Sullivan, S., Chittka, L., 2009. Bird pollination of Canary Island endemic plants. Naturwissenschaften 96, 221-232. Ortega-Olivencia, A., Rodriguez-Riano, T., Valtuena, F.J., Lopez, J., Devesa, J.A., 2005. First confirmation of a native bird-pollinated plant in Europe. Oikos 110, 578-590. Percival, M., 1965. Floral biology. Pergamon Press, Oxford. Perret, M., Chautems, A., Spichiger, R., Peixoto, M., Saudaiven, V., 2001. Nectar sugar composition in relation to pollination syndromes in Sinningieae (Gesneriaceae). Annals of Botany 87, 267-273. Peters, W.S., Pirl, M., Gottsberger, G., Peters, D.S., 1995. Pollination of the Crown Imperial Fritillaria imperialis by great tits Parus major. Journal für Ornithologie 136, 207-212. Pike, G.H., Waser, N.M., 1981. The production of diluted nectar by hummingbird and honeyeater flowers. Biotropica 13, 260-270. Proctor, M., Yeo, P., 1973. The pollination of flowers. Collins, London. Proctor, M., Yeo, P., Lack, A., 1996. The natural history of pollination. Timber Press, Portland. Ramsey, J., Bradshaw, H.D.J., Schemske, W., 2003. Components of reproductive isolation between the monkeyflowers Mimulus lewisii and Mimulus cardinalis (Phrymaceae). Evolution 57, 1520-1534. Raven, P.H., 1972. Why are red-visited flowers predominantly red? . Evolution 26, 674. Rodriguez-Gironés, M.A., Santamaría, L., 2004. Why are so many bird flowers red? . PLoS Biology 2, 350. 37  Schemske, D.W., Bradshaw, H.D.J., 1999. Pollinator preference and the evolution of floral traits in monkeyflowers (Mimulus) Proceedings of the National Academy of Sciences of the United States of America 96, 11910-11915. Schuler, W., 1977. Birds and sugars: starling prefers glucose and fructose but reject sucrose. Abstract XVII International Ornithological Congress. Schwilch, R., Mantovani, R., Spina, F., Jenni, L., 2001. Nectar consumption of warblers after long-distance flights during spring migration. Ibis 143, 24-32. Scogin, R., 1988. Floral anthocyanidins of bird-visited flowers. Botanical Gazette 149, 437-442. Snow, B.K., Snow, D.W., 1972. Feeding niches of hummingbirds in a Trinidad Valley. Journal of Animal Ecology 41, 471-485. Specht, C.D., 2006. Systematic and evolution of Costaceae (Zingiberales): a multiple dataset approach. Systematic Botany 31, 89-106. Stelzer, R., 2005. Sammelstrategien bei Hummeln ein Vergleich zwischen Insel und Festlandpopulationen. Fakultät für Biologie. Universität Würzburg. Stiles, F.G., 1975. Ecology, flowering phenology, and hummingburds pollination of some Costa Rican Heliconia species. Ecology 56, 285-301. Stiles, F.G., 1978. Ecological and evolutionary implications of bird pollination. American Zoologist 18, 715-727. Stiles, F.G., 1981. Geographical aspects of bird-flower coevolution, with particular reference to Central America. Annals of the Missouri Botanical Garden 68, 323-351. Stiles, F.G., Freeman, C.E., 1993. Patterns in floral nectar characteristics of some bird-visited plant species from Costa Rica. Biotropica 25, 191-205.  38  Thomson, J.D., Wilson, P., Valenzuela, M., Malzone, M., 2000. Pollen presentation and pollination syndromes, with special reference to Penstemon. Plant Species Biology 15, 1129. Toledo, V.M., 1975. Chiranthodendron pentadactylon Aristegnii (Sterculiaceae): una especie polinizada por aves percheras. Boletín de la Sociedad Botanica de Mexico 35, 42-52. Valido, A., Dupont, Y.L., Olesen, J.M., 2004. Bird-flower interactions in the Macaronesian islands. Journal of Biogeography 31, 1945-1953. Van der Pijl, L., Dodson, C.H., 1966. Orchid flowers, their pollinators and evolution. University of Miami Press, Coral Gables. Waser, N.M., Chittka, L., Price, M.V., Williams, N.M., Ollerton, J., 1996. Generalization in pollination systems, and why it matters. Ecology 77, 1043-1060. White, G., 1789. The Natural History and Antiquities of Selborne, London. Yanega, G.M., Rubega, M.A., 2004. Feeding mechanisms: Hummingbird jaw bends to aid insect capture. Nature 428, 615. Zufall, R.A., Rausher, M.D., 2003. The genetic basis of a flower color polymorphism in the common morning glory (Ipomoea purpurea). Journal of Heredity 94, 442-448. Zufall, R.A., Rausher, M.D., 2004. Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428, 847-850.  39  3 Evolution of petal epidermal micromorphology in Leguminosae and its use as a marker of petal identity1  3.1 Introduction The petal epidermal surface is important in pollination, as it influences the way in which pollinators perceive and interact with the flower. There is evidence that petal epidermal type and its surface affect colour depth (Dyer et al., 2007; Gorton and Vogelmann, 1996; Kay, 1988; Kay et al., 1981), iridescence (Whitney et al., 2009b), scent production (Kolossova et al., 2001), temperature (Comba et al., 2000; Dyer et al., 2006) and provides tactile cues (Comba et al., 2000; Kevan and Lane, 1985b; Stirton, 1981; Whitney et al., 2009b). Previous surveys have classified and analyzed the distribution of the epidermal surface of petals within the angiosperms (Barthlott and Ehler, 1977; Kay et al., 1981). Extensive and detailed analyses have been provided for a few groups, such as Asteraceae (Baagøe, 1977, 1980; Hansen, 1991) and Leguminosae (Stirton, 1981), in which the characteristics of these epidermal types have been used for taxonomic and phylogenetic analyses. Among the epidermal types described in these studies, papillose conical cells are frequently reported. Between 60 to 80 % of the angiosperm species analyzed have at least one petal with this epidermal type on the adaxial surface (or upper side, towards the floral axis) of the petals (Christensen and Hansen, 1998; Kay et al., 1981). The frequency of this cell type, coupled with other evidence, suggests that it has an adaptive value (Glover and Martin, 2002).  1  A version of this chapter has been published. Ojeda I, Francisco-Ortega, J. and Cronk, Q. (2009)  Evolution of petal epidermal micromorphology in Leguminosae and its use as a marker of petal identity. Annals of Botany 104 (6): 1099-1110. 40  Several recent studies have elucidated the genetic basis of the papillose conical cell type. The MIXTA gene, a MYB transcription factor, is required for the development of papillose conical cells in Antirrhinum majus L. (Noda et al., 1994). Further evidence suggests that this gene is sufficient to produce papillose conical cells in other groups within the asterid clade (Solanaceae and Scrophulariaceae). The developmental program leading to this epidermal pattern is therefore quite conserved, at least in these groups (Glover and Martin, 2002). However, it seems that a different developmental pathway may operate in the Brassicaceae (rosid clade) (Glover et al., 1998; Payne et al., 1998). Petal identity, whether dorsal (adaxial), lateral or ventral (abaxial) is set at an early developmental stage by the expression of identity genes that induce transcription of other genes responsible for the shape, size, colour and epidermal type characteristics of individual petals (Cronk, 2006a). A change of petal identity may therefore change features of the epidermal surface. The functions of two petal identity genes have so far been described in the Leguminosae (Feng et al., 2006b). One of these genes, Lotus japonicus CYCLOIDEA 2 (LjCyc2) is a transcription factor in the TCP gene family (named after its first three characterised members, TB1 in maize, CYC in snapdragon and PCFs in rice) with a dorsalizing action. This gene is associated with dorsal petal identity and hence papillose conical cell formation, as papillose conical cells are characteristic of standard petals in L. japonicus. The second gene, LjCYC3 is a lateralizing factor. This gene is responsible for the lateral petal identity and hence the formation of tabular rugose cells with a jigsaw puzzle-shape that are characteristic of wing petals. The molecular basis of jigsaw puzzle shape cells has recently been elucidated (Fu et al., 2005). These results suggest that each petal in the papilionoid legume flower has a distinct molecular identity, that may be marked by epidermal type. In transgenic plants that overexpress LjCYC2 all petals have papillose conical cells, indicating that all petals have been converted to a 41  dorsal identity. Furthermore, natural evolution in legume flowers may occur by shifts in petal identity (Citerne et al., 2006). Many legume species have a specialized flower morphology that promotes pollinator specificity (Fig. 3.1). Papilionoid legumes generally have three kinds of petals: a dorsal petal, called a "standard", two lateral petals or "wings", and two ventral petals forming the "keel". The distribution of the epidermal types within this family has not been analyzed in detail. A few studies have included some legume species (Christensen and Hansen, 1998; Kay et al., 1981), but in the first case only the dorsal petal was analyzed (C.I. Christensen, University of Copenhagen, Denmark, pers. comm.). A more extensive survey (Stirton, 1981) focused only on the papilionoids and was restricted to lateral petals. Another study has reported the epidermal type of two Lathyrus species, but only the dorsal and lateral petals were analyzed (Hammett et al., 1994). There is therefore a lack of detailed information about the epidermal types within the family. Knowledge of the molecular developmental basis of petal identity has therefore not been coupled with a systematic analysis of the epidermal types and their distribution within the Leguminosae, even though epidermal types could be useful micromorphological markers of petal identity in developmental studies, for instance in the analysis of developmental mutants. In order to address this lack of information, we conducted a survey of the epidermal surface of the various types of petals within this family.  3.1.1 Objectives of the study The aims of the present study are (1) to characterize the epidermal cell types of petals in the major clades of the Leguminosae, (2) to determine their distribution along four axes of distribution, one within the flower (dorsiventral) and three within each petal (abaxial-adaxial 42  surfaces, proximo-distal and medio-lateral), and (3) to relate these patterns to our current understanding of the molecular genetic basis of petal identity and flower evolution within the family.  3.2 Material and methods 3.2.1 Taxon sampling Representative species were chosen from each of the three subfamilies (Caesalpinioideae, Mimosoideae and Papilionoideae) and all 12 major clades currently recognized in the Leguminosae following recent phylogenetic analyses and taxonomic treatments (Lavin et al., 2005; Lewis et al., 2005; Wojciechowski et al., 2004). In total our sampling included 175 species representing 26 of the 37 tribes of the family, and 89 genera (Table 3.1) (Lewis et al., 2005). For an outgroup comparison four species of the closely related family Polygalaceae were analyzed. Polygalaceae also has zygomorphic flowers, but of a different constitution (Prenner, 2004a; Prenner and Klitgaard, 2008). Comparison with Polygalaceae is problematic as in this family the two adaxial petaloid organs are sepals. The lateral petals are reduced and the two adaxial petals may be considered to have closest functional equivalence to the lateral petals in legumes. The single abaxial petal forms the keel-like structure. We have therefore chosen to compare functionally equivalent structures (Table 3.2). Mimosoid legume flowers also have a contrasting morphology (Prenner, 2004b) with an abaxial median petal rather than an adaxial median petal. In this case the dorsal petals are compared to the dorsal petal in other legumes and the ventral petal is compared to the ventral petals in other legumes. To explore the variation within species several individuals were studied in two taxa (Lotus japonicus, four individuals and Trifolium repens, three individuals) but subsequent 43  sampling used only one individual to represent each species as intra-specific variation was found to be negligible. The variation among species within the same genus was explored in 11 genera, in all of which four or more species were analyzed. In Senna, 12 species were included representing three of the six recognized sections (Chamaefistula, Senna, Peiranisia) (Irwin and Barneby, 1982). These species represent all major clades currently recognized in this genus and included the entire range of flower symmetry within the group (Marazzi and Endress, 2008; Marazzi et al., 2006). Seven species of Lathyrus belonging to two of the nine sections of this genus (Lathyrus and Orobus) were studied (Kenicer et al., 2005). In Lotus seven species from three of its 14 sections (Bonjeana, Lotus and Pedrosia) were included (Degtjareva et al., 2006). In Bauhinia we analyzed six species with diverse flower morphology. In Cassia six species were included with a range of flower morphology and petal differentiation. Six species of Dalbergia were analyzed. Within Erythrina five species with diverse flower morphology were included. Within Vicia five species were sampled representing the two subgenera Cracca (sections Cracca and Cassubicae) and Vicia (sections Vicia and Faba) currently recognized within this genus (Choi et al., 2006a). In Trifolium, five species belonging to two of its subgenera, Chronosemium and Trifolium, and including sections Trifoliastrum, Involucrarium, Trifolium and Vesicastrum, were analyzed (Ellison et al., 2006). In Genista we sampled four species representating the subgenera Spartocarpus (section Spartocarpus) and Genista (section Spartioides), together with the species Ulex europaeus and Chamaespartium sagittale, currently considered to be nested within this group (Pardo et al., 2004). Finally, four species of Dalea were included.  3.2.2 Microscopy and cell type classification Petals of fully open mature flowers were the subject of micromorphological studies using either a Hitachi S-2600N or a JEOL JSM-5900 LV scanning electron microscope (SEM) at an 44  acceleration voltage of 10-15 Kv. Either high vacuum or low vacuum conditions were used without significant differences on the structures observed. Some species were analyzed using a light microscope (Motic B Series) from preserved flowers in 70% ethanol or from herbarium specimens. In the latter case the flowers were first rehydrated, fixed in FAA, preserved in 70% ethanol and then analyzed. One species, Lotus japonicus, was tested under all different conditions to determine treatment effects or artifacts. Entire or partial petals were mounted on SEM stubs with double-sided adhesive tape. The distribution of the epidermal types was recorded using the following four axes using as a reference the centre of the flower (Fig. 3.2,AF),where differentiation may occur, as listed below: 1) a dorsiventral axis within the flower. In the case of zygomorphic (monosymmetric) flowers (Endress, 2001), the dorsal, one lateral and one ventral petal were analyzed (Fig. 3.2,A, B). In actinomorphic (or polysymmetric) flowers all five petals were analyzed. In some caesalpinioids, with asymmetric flowers, all five petals were also analyzed separately. 2) abaxial-adaxial axis within the petal, i.e. upper and lower surfaces. 3) a proximo-distal axis within each petal, i.e. base to tip (Fig. 3.2, D-F). 4) a medio-lateral axis within the petal, i.e. middle to edge (Fig. 3.2, D-F]. The epidermal types were classified based on cell-shape traits (the primary sculpture) and on the fine relief of the cell wall (or secondary sculpture) (Barthlott, 1981, 1990), using the standard terminology regularly applied in similar studies within angiosperms (Kay et al., 1981). The epidermal types were classified into two main types: papillose and tabular (Table 3.3). These two types were further subdivided based on cell shape and sculpture of the outer surface. The distribution of these epidermal types was then recorded in relation to the four axes described above, either on entire petals (in the case of small flowers) or portions (representing each one of the sections) in the case of species with large flowers. 45  3.2.3 Reconstruction of evolutionary changes Character evolution was studied using parsimony (DELTRAN) as implemented in MacClade 4.0 (Maddison and Maddison, 2000) and maximum likelihood using Mesquite (Maddison and Maddison, 2009). No major differences were found using the different methods. Traits were coded as binary characters (absence and presence). The distribution of these traits was mapped on a tree built according to recent phylogenetic analyses of the legumes (Lavin et al., 2005; Wojciechowski et al., 2004). Additional phylogenetic studies of individual genera or tribes were consulted in order to have a fully resolved phylogeny (Choi et al., 2006a; Degtjareva et al., 2006; Ellison et al., 2006; Kenicer et al., 2005; Marazzi et al., 2006; McMahon and Hufford, 2004; Pardo et al., 2004). In a few cases where there was a lack of information on specific groups, the polytomies were resolved randomly using MacClade.  3.3. Results 3.3.1 Epidermal types within Leguminosae A total of six major epidermal types were recorded; five of which occurred in the papilionoids, three were found in the caesalpinioids and only one type was detected in the mimosoids (Table 3.3) (Fig. 3.3, A-R). Of the major categories only the tabular rugose cells with striations (TRS) had marked variation in cell shape, size and in the fine relief of the cell wall. Therefore, this epidermal type was further subdivided into three minor subtypes (i, ii and iii) described in Table 3.3. Some papilionoid species had only TRS on all three types of petals, but these minor differences (as marked by the TRS subtypes) can potentially be used to distinguish petal types in at least some species (Table 3.3 and Fig. 3.4).  46  3.3.2 Strong micromorphological variation in petals is exclusive to papilionoid legumes The three distinct types of petals found in papilionoids generally showed clear micromorphological differences at the epidermal level (Fig. 3.5, A). Character analysis using parsimony and maximum likelihood indicates that this represents a single character state gain (Fig. 3.6). The early divergent Cladrastis branch of the papilionoids has only two major epidermal types, one occurring on the dorsal and lateral petals, and the other found on the ventral petal. Within papilionoids two groups, the tribe Loteae and the genistoid clade, commonly have the greatest micromorphological differentiation; their standards, wings, and keels are each characterized by specific epidermal features. Similarly, all of the five epidermal types that occur in papilionoids tend to be associated with a particular petal type (Table 3.4). For instance tabular flat striate cells (TFS) are restricted to keel petals and papillose conical cells (PCS) are generally characteristic of the standard petal. Further details of the association between epidermal types and petal types are given elsewhere (Table 3.2). An interesting feature of papilionoid legumes is the occasional presence of more than one epidermal type on the same petal. In general the base of each petal has poorly differentiated cells (cells of simple shape, without prominent surface features and characteristics of early developmental stages), while cells in the middle and distal part are more strongly differentiated (Fig. 3.7,A-J). However, in some species two strongly differentiated but quite different epidermal types were observed in the same petal (Fig. 3.7). For example, in Lathyrus the dorsal petal has mainly tabular rugose cells (TRS) but has a border with papillose lobular cells (PLS) (Fig. 3.7,K-P).  47  3.3.4 Loss of micromorphological variation in Indigofera, Amorpheae and IRLC. Although most papilionoids have high micromorphological variation along their floral dorsiventral axis, three papilionoid groups do not exhibit this pattern and have only one major epidermal type on all petals. The Indigofera clade, the Amorpheae and most species of the large Inverted Repeat-lacking clade (IRLC) have TRS covering most of their petals. In these groups, papillose conical cells are absent from the dorsal petal (although there are numerous exceptions, as noted in Table 3.2) (Fig. 3.6) and the area covered by TFS on the ventral petal is commonly reduced to a small region at the tip. In the Amorpheae, all species analyzed have TRS on the three types of petals, with the exception of Dalea leporina, which has TRS on ventral and PCS on dorsal and lateral petals. The lack of micromorphological variation is evident both in species with a typical papilionaceous corolla, with three types of petals, such as in Psorothamnus arborescens and Marina spp. and in species with two petal types, such as Apoplanesia paniculata. It is noteworthy that although TRS is the dominant epidermal type in all three petals, the ventral petal of these groups may still have a small amount of the TFS epidermal type that characterizes this petal type in other papilionoids. Furthermore, although these groups have the same major epidermal type (TRS) on the dorsal petal and on the wings, in some cases we detected minor differences in cell size and shape that made it possible to distinguish them easily. This was particularly clear in the IRLC. Most of the species in this clade have lost the diversity of major epidermal types, as they have only tabular rugose cells with striations (TRS); however, in some instances we found striking differences between the TRS cell morphology on different petals within the flower. Variation in several features, such as cell size, shape of the cells (whether elongated or isodiametric), marginal features (such as the waviness of the cell  48  margin) and features of the surface (such as density of striations), allows clear micromorphological identification of each petal type. I therefore further subdivided the TRS cell type into three subtypes: TRS subtype i (TRSi) has elongated cells, TRS subtype ii (TRSii) has isodiametric cells. In these two subtypes, the cell wall is usually well delimitated with dense striations. In contrast, TRS subtype iii (TRSiii) has elongated cells that tend to have weak cell wall delimitation and a lower density of striations. TRSi and TRSii were mainly observed on the dorsal and lateral petals, respectively, while TRSiii was found exclusively in the ventral petal (Fig. 3.4). In some of the IRLC species, e.g. Pisum sativum (Fig. 3.4, A-C), this variation allows all three types of petals to be differentiated, or at least for the ventral (keel petal) to be differentiated from the dorsal and lateral petals, as in Vicia hirsuta (Fig. 3.4, D-F) or Trifolium repens (Fig. 3.4, G-I). However, in some IRLC species, such as Melilotus spp. (Fig. 3.4, J-L), there are no obvious differences between the different petals (Table 3.2).  3.3.5 Zygomorphy in caesalpinioids is not associated with strong micromorphological variation The caesalpinioids show micromorphological differences of major epidermal types between species (Fig. 3.5, B-D and Fig. 3.4, M-O). Despite the variation in petal size and shape within flowers of many of the caesalpinioids surveyed (Fig. 3.2, A-F), the major epidermal type of each petal within a flower is uniform, usually TRS (64 %) while PCS is less common. Character state reconstruction indicates that TRS is the ancestral petal cell type in caesalpinioids and in legumes as a whole (Fig. 3.6). There is therefore no association of a particular major epidermal type with any specific petal, although we found some very minor variations in cell size and form. Even Cercis, a 49  caesalpinioid with strongly zygomorphic flowers that superficially resemble papilionoid flowers has all its petals as TRS with only minor differences distinguishing them (Fig. 3.4, M-O). We included species representing all the main variation of flower symmetry within the genus Senna, which ranges from radial to asymmetric flowers (enantiostyly) (Marazzi and Endress, 2008; Marazzi et al., 2006). All the petals in individual flowers of this genus have the same major epidermal type (either TRS, PCS or PKR), even in species such as Senna mucronifera or Cassia emarginata (Fig. 3.2, E), where each of the five petals are different in size and shape.  3.3.6 The occurrence of papillose cells in the Leguminosae Papillose cells (of all types) are particularly characteristic of papilionoids, but also appear to have evolved independently in some caesalpinioids (Fig. 3.6). Papillose cells (PCS and PKR) are mainly found in the dorsal part of the papilionoid flower (especially the standard). However some species have papillose cells on the wings. More rarely, papillose cells are found on the keel. For example, in Lespedeza thunbergii (Fig. 3.2, H) and Desmodium incanum, the keel petals are mainly covered by TFS, as in the majority of papilionoids. However, the tip of the keel has some cells intermediate between tabular flat cells (TFS) and papillose conical cells (PCS) in the most exposed area. In the 175 species analyzed in this study only seven species (Dalbergia brownei, Canavalia rosea and five species of Erythrina) had well developed papillose conical cells on the keel. Dalbergia brownei and Canavalia rosea display PCS only in small patches. The genus Erythrina is notable for having papillose cells (PCS and PKR) on all three types of petal.  50  3.4 Discussion 3.4.1 Functional significance of papillose cell types Papillose cells are characteristic of many papilionoid lineages and this group of cell types may have evolved in papilionoids due to possible functional advantages. Papillose cells may increase petal brightness and therefore increase pollinator visitation rates (Comba et al., 2000; Dyer et al., 2006; Glover and Martin, 1998). It is interesting to consider why papillose cells (PCS and PKR) tend to be characteristic of the dorsal and lateral petals in papilionoids (Fig. 3.5, A), while these cell types are virtually absent in the ventral petals of the papilionoid subfamily (Table 3.4). Functionally, this may be because keel petals are in many cases covered by the lateral petals and are not usually prominent in pollinator attraction. This explanation is supported by the fact that species with papillose cells on the keel generally have an exposed keel, which probably does function in pollinator attraction (e.g. Erythrina, Crotalaria and Lespedeza).  3.4.2 Lack of micromorphological variation in the Caesalpinioideae In contrast to papilionoids, caesalpinioids have relatively little micromorphological variation. Differences between dorsal, lateral and ventral petals do occur but are small and never involve major epidermal types (Fig. 3.5). This implies that a strong connection between the dorsiventral patterning of the flower and the developmental patterning of cell form, that is so evident in papilionoids, never evolved in caesalpinioids. Within the legumes, this connection between symmetry and cell type differentiation is therefore a unique derived character of papilionoids. This lack of variation in caesalpinioids is independent of floral patterning, from nearly radial, as in Bauhinia natalensis and B. petersiana, to strongly zygomorphic flowers, as in Tara 51  cacalao or Cercis spp. (Fig. 3.2, D). For instance, Cercis has highly zygomorphic flowers (Fig. 3.2, A), and still displays much less micromorphological variation (Fig. 3.4, M-O) than a typical papilionoid (Fig. 3.5, A). These results support previous studies suggesting that the floral morphology of this genus, although superficially similar to a papilionoid, is only a weak convergence at the anatomical level to the papilionoid flower (Tucker, 2002). In most caesalpinioids the adaxial and abaxial surface of the petal have the same major epidermal type. However, this is not the case in Senna alata (with all petals having PCS on the abaxial side and PKR on the adaxial one) and Bauhinia tomentosa (with PCS on the abaxial side and TRS on the adaxial side of all petals) (Table 3.2). These two species have flowers which do not fully open and it is the abaxial (exposed) (Fig. 3.2, B) surface that has papillose cells, which may enhance brightness and therefore the pollinator attractiveness of the flowers. Another interesting feature of the caesalpinioids is the presence of trichomes in about 29% (11 species of 39) of the species analyzed and stomata in 15% (6/39). In most species the distribution of trichomes was homogeneous on all the petals within the flower (all petals having trichomes). This feature is shown in Cassia emarginata (Fig. 3.5). However, in six species trichomes were localized on a specific petal and hence the distribution of the trichomes could be used as an indicator of petal identity in these species. In Table 3.2 species with trichomes and stomata are denoted by a superscript t and s, respectively. Trichomes and stomata are also occasionally found in papilionoids but more rarely.  3.4.3 Mixing of epidermal types in the same petal surface A unique feature of papilionoids is the occasional occurrence of more than one major epidermal type within a single petal. Between the two major epidermal types there is a transition zone or morphocline (Baagøe, 1977; Hansen, 1991), which has been reported previously in other 52  groups of angiosperms (Barthlott and Ehler, 1977; Christensen and Hansen, 1998; Hansen, 1991, 1992). The shifts observed involve changes from TRS to PLS (Fig. 3.7, K-P), TRS to PCS (Fig. 3.7, Q-V) or TRS to PKR in the dorsal and lateral petals, and from TRS to TFS in the ventral petal. The transition zone, with intermediate cell morphology between the two major epidermal types, is always relatively narrow (Fig. 3.7). If the epidermal cell type is responding to the expression of underlying petal identity genes, as seems to be the case in at least some legumes (Feng et al., 2006b; Wang et al., 2008), these morphoclines may then indicate gene expression boundaries within organs. It might even suggest that such petals are of "mixed identity" and hence developmentally composite organs.  3.4.4 Contrasting loss of micromorphological variation within flowers of the Amorpheae and IRLC Despite the striking micromorphological variation observed within the papilionoid flower, some species lack this variation. This is especially evident in the Amorpheae. It is interesting to note that the Amorpheae is often characterized by flowers of altered zygomorphy. They vary from zygomorphic (flowers with corollas of three types of petals as in Psorothamnus, Marina and some Dalea species) to subactinomorphic (five petals poorly differentiated into two types as in Apoplanesia and Eysenhardtia), while some species have only one petal (Amorpha) and other species lack petals altogether (Errazurizia and Parryella) (McMahon, 2005; McMahon and Hufford, 2004, 2005). The tendency to weak dorsiventrality and subactinomorphy in this group is therefore associated with a lack of micromorphological diversity in the major epidermal types. Other dalbergioids, the group to which Amorpheae belongs, generally have flowers typical of other papilionoids. 53  The IRLC has also lost diversity of major epidermal types (TRS being the most common type), but in this case it is not accompanied by any loss of dorsiventral patterning as the flowers are highly zygomorphic and some species show differentiation in subtypes of TRS (Fig. 3.4). This indicates that the loss of epidermal diversity in Amorpheae and IRLC may be different in mechanism with different underlying genetic control in each group.  3.4.5 Genetic control of petal micromorphology and petal identity Most studies of genetic control of petal micromorphology have focused on papillose conical cells. MIXTA, a transcription factor of the MYB family, has been associated with the differentiation of PCS in Anthirrinum majus, and in some species of Solanaceae (Glover and Martin, 2002; Noda et al., 1994). In Petunia hybrida Vilm. an ortholog of MIXTA, mybPh1, is associated with conical cells (Avila et al., 1993; van Houwelingen et al., 1998). However, there is not yet any evidence that MIXTA homologues play a role in PCS differentiation in legumes. An additional network of genes has been associated with the jigsaw puzzle shape of Arabidopsis leaf pavement cells. Pavement cell morphogenesis is controlled by two antagonistic pathways, Rho GTPase (ROP) and ROP effector protein (RIC) pairs with opposing action. The countersignaling of these two pathways has been associated with the interdigitation and final shape observed in these epidermal cells, with lobes and indentations. The pair ROP2/RIC4 promotes cell growth on lobes, and the gene pair ROP2/RIC1 restrains outgrowth, hence producing the indentations (Fu et al., 2005; Guimil and Dunand, 2007; Mathur, 2006). There is at present no information as the precise pathway by which petal identity genes promote the differentiation of the epidermal types described in this survey. Lotus japonicus CYCLOIDEA 2 (LjCYC2), a transcription factor of the TCP family, promotes dorsal petal identity, i.e. the expression of this gene activates other gene networks 54  necessary for dorsal petal traits and it therefore confers specific organ fate on the primordium in which it is expressed. LjCYC2 expression is therefore necessary for the differentiation of PCS in this petal (Feng et al., 2006b). Overexpression of this gene in transgenic plants promotes a dorsalization of all petals, with the consequent production of PCS cells in all petals. In Lotus japonicus, LjCYC2 is exclusively expressed on the adaxial side of the flower (dorsal petal) and thus affects dorsal petal identity (Feng et al., 2006b). It has been demonstrated that in species with a gain of dorsal identity, such as Cadia purpurea (G. Piccioli) Aiton, this gene is expressed throughout the flower, and all petals have the same shape, symmetry and identity (Citerne et al., 2006). Studies of the petal micromorphology of Cadia and related species would therefore be of great interest. Another gene, Lotus japonicus KEELED WING 1 (KEW1 or LjCYC3), is associated with lateral petal identity. LjCYC3 is required for normal lateral petal development, which includes tabular rugose cells (TRS) with a jigsaw puzzle shape. As in LjCYC2, mutations that knockout the activity of LjCYC3 causes a ventralization of the lateral petal, with a subsequent lack of TRS and the presence of TFS (Feng et al., 2006b). To date, the role of LjCYC2 has been only explored in detail in Lotus japonicus (Feng et al., 2006b) and Pisum sativum (Wang et al., 2008). Orthologues of LjCYC2 (PsCYC2) and of LjCYC3 (PsCYC3) have recently been cloned in Pisum sativum (Wang et al., 2008). These two genes are required for normal zygomorphic development. PsCYC2 is associated with dorsal identity and PsCYC3 with lateral identity. But unlike L. japonicus, all petals have the same major micromorphology (TRS) and the identity of each petal (dorsal, lateral and ventral) is associated with the variation of features within this major epidermal type (Wang et al., 2008). Homologues of CYCLOIDEA have been also explored in Genisteae (Citerne et al., 2003; Citerne et al., 2006; Ree et al., 2004). 55  Therefore, the activation of the ROP/RIC pathway, essential for jigsaw puzzle cell shape, must be downstream of the LjCYC3 identity gene. No ventral identity gene has yet been found, and the cause of ventral petal identity (and hence TFS differentiation) in legumes is still unknown. However, in both Lotus japonicus and Pisum sativum, the double knockout of the dorsal and lateral identity genes causes a ventralization of all the petals, suggesting that ventral identity and TFS are perhaps the default states in these groups.  3.4.6 Evolution of petal micromorphology I have have shown that in the Caesalpiniaceae different petals within a flower are not strongly distinguished micromorphologically and the most common pattern is TRS/TRS/TRS (in dorsal, lateral and ventral petals respectively). The ancestral state reconstruction analysis suggests that this pattern probably represents the ancestral condition in the Leguminosae (Fig. 3.6). More or less papillose cells have independently evolved at least six times in the Caesalpinioideae and, when they occur, species have either PKR (in all petals) or PCS (in all the petals). Papillose cells appear to be absent in Mimosoideae, and their very small petals all have TRG. In addition, my study suggests that the strong micromorphological differentiation within the flower is an advanced condition of the family that has evolved only within the papilionoid clade. Dorsiventral differentiation of major epidermal types appears to have evolved at the base of the papilionoid clade (with the less derived papilionoids state being PKR/PKR/TFS), reaching maximum differentiation independently in Loteae (PCS/TRS/TFS) and genistoids (PKR/PCS/TFS). Papillose cell types appear to have evolved many times in legumes but may have evolved only once in papilionoid legumes (at the base of that clade) and this character has apparently been lost at least four times in the subfamily (Fig. 3.6). However, most of the 56  papilionoid lineages that have lost papillose cells (groups predominantly with TRS/TRS/TRS as a character reversal), still display dorsiventral differentiation between petals by means of different TRS subtypes (e.g. TRSi, TRSii, and TRSiii).  Table 3.1 List of species sampled during this study. Clades are recognized following Wojciechowski et al., 2004; Lavin et al., 2005; Lewis et al., 2005. Tribal classification following Lewis et al., 2005. * = species analyzed from fresh flowers preserved in ethanol using a light microscope, + = species analyzed from flowers re-dehydrated from voucher specimens and preserved in ethanol using a light microscope. Other species were studied using the SEM and fresh material. UBCBG = UBC Botanical Garden, FTBG= Fairchild Tropical Botanical Garden, MBG = Montgomery Botanical Garden, QEG = Queen Elizabeth Garden, Vancouver, JBRCICY = Jardín Botánico regional del CICY, CICY = Centro de Investigación Científica de Yucatán, Mexico. JAO = Jardín de Aclimatación de la Orotova, Spain.  57  Taxon  Collection site and number  Voucher information  POLYGALACEAE Polygala chamaebuxus L. P. diversifolia L. P. grandiflora Walter P. myrtifolia L.  UBCBG, No. 31175-575-94 FTBG, No. 941312A Wild collected, Big Pine Key, Florida FTBG, No. 2007-0032A  Ojeda 56 Ojeda 111 Ojeda 121 Ojeda 116  UBCBG, No. 03872-0691-2006 UBCBG, No. 27814-568-89 MBG, No. S-19-04 MBG, No. 2006-0207 FTBG, No. 2002-0513A FTBG, No. 981198A MBG, No. S-67-06 JAO, No. 348-99 FTBG, No. 941095B  Ojeda 48 Ojeda 47 Ojeda 92 Ojeda 95 Ojeda 127 Ojeda 105 Ojeda 90 Ojeda 161 Ojeda 112  JAO, No. 0270-70 FTBG, No. 821821D Public street, Coral Gables, Florida  Ojeda 155 Ojeda 104 Ojeda 130  JAO  Ojeda 159  FTBG, No. 2000-119D JAO, No. 14-96 MBG, No. 7687A MBG, No. 91268A Cultivated, Australia Queesland FTBG, No. 7978A FTBG, No. F64901A Wild collected, Big Pine Key, Florida FTBG, No. 2006-1082A Blodel conservatory QEG FTBG, No. 2006-1082A MBG, without number Wild collected, Paraguay, Caaguazú  Ojeda 126 Ojeda 156 Ojeda 101 Ojeda 87 Endress 6411 Ojeda 128 Ojeda 113 Ojeda 122 Ojeda 107 Ojeda 40 Ojeda 109 Ojeda 131 Marazzi et al. BM 019 Marazzi et al. BM 185 Marazzi et al. BM 178 Ojeda 93 Ojeda 148 Ojeda 96  LEGUMINOSAE CAESALPINOIDEAE Cercis crown node Tribe Cercidae Cercis canadensis L. C. yunnanensis Hu & Cheng Bauhinia divaricata L. B. natalensis Oliv. B. petersiana Bolle B. tarapotensis Benth. B. tomentosa L. *B. variegata L. Tylosema fassoglensis (Kotschy ex Schweinf.) Torre & Hillc. Caesalpioid crown node Tribe Detarieae *Brownea ariza Benth. B. capitella Jacq. Tamarindus indica L. Umtiza crown node Tribe Caesalpineae *Gleditsia triacanthos L. Tribe Cassieae Cassia emarginata L. *C. grandis (L.F.) Pers. C. fistula L. C. javanica L. *C. javanica L. C. nodosa Buch.-Ham. ex Roxb. C. roxburghii DC. Chamaecrista lineata (Sw.) Greene Senna alata (L.) Roxb. S. corymbosa (Lam.) H.S. Irwin & Barneby S. ligustrina (L.) H.S. Irwin & Barneby S. mexicana (Jacq.) H.S. Irwin & Barneby *S. mucronifera (Mart. ex Benth.) H.S. Irwin & Barneby *S. nicaraguensis (Benth.) H.S. Irwin & Barneby *S. pallida (Vahl) H.S. Irwin & Barneby  Wild collected, México, Chiapas Wild collected, México, Oaxaca  S. polyphylla (Jacq.) H.S. Irwin & Barneby *S. racemosa (Mill.) H.S. Irwin & Barneby S. septemtrionalis (Viv.) H.S. Irwin & Barneby *S. siamea (Lam.) H.S. Irwin & Barneby  MBG, No. S-111-05 JBRCICY MBG, No. 2002-0775  *S. wislizeni (A. Gray) H.S. Irwin & Barneby  Wild collected, México, Puebla  Tribe Caesalpineae Caesalpinia pulcherrima (L.) Sw. *C. gaumeri Greenm.  FTBG, No. 2003-0464A JBRCICY  Wild collected, Panamá City, Panamá  Marazzi et al. BM 157 Marazzi et al. BM 169 Ojeda 114 Ojeda 150  58  Taxon  Collection site and number  Delonix regia (Bojer ex Hook.) Raf. Peltophorum pterocarpum (DC.) Backer ex K. Heyne Tara cacalao unpub. combination  MBG, No. 2000-1152A MBG, No. 76178A  Voucher information Ojeda 103 Ojeda 83  FTBG, No. 93585C  Ojeda 117  FTBG, No. 67330C  Ojeda 129  Public street, Vancouver Blodel Conservatory QEG MBG, No. 2002-0124A MBG, No. 99568A MBG, No. 2003-1080B  Ojeda 42 Ojeda 41 Ojeda 100 Ojeda 86 Ojeda 102  MBG, without number  Ojeda 88  PAPILIONOIDEAE Swartzia crown node Tribe Swartzieae +Swartzia pittieri Schery  Voucher specimen, CICY herbarium  G. Carnevali & F. Guanhez 1491  Cladrastis crown node [Tribe Sophoreae] +Cladrastis lutea (Michx.) K. Koch +C. sinensis Hemsl.  Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium  Jenninson 558 Straley 4284  Public garden, Vaca Key, Florida FTBG, No. 2006-0220A  Ojeda 119 Ojeda 108  UBCBG, No. 32873-620-95 UBCBG, No. 037228-0474-2004 UBCBG UBCBG, No. 005631-0171 UBC campus, Vancouver UBCBG, No. 10584-156-74 UBCBG, No. 037943-0117-2005 UBCBG, No. 0112128-0117-2005 UBCBG, No. 036849-5491-2003 UBCBG, No. 22652-050-82 UBCBG, No. 023189-0284-1983 Spanish Bank beach, Vancouver UBC campus, Vancouver Wild collected, Tenerife, Spain UBCBG, No. 014629-0275-1977 UBCBG, without number  Ojeda 50 Ojeda 61 Ojeda 62 Ojeda 20 Ojeda 51 Ojeda 55 Ojeda 49 Ojeda 54 Ojeda 134 Ojeda 57 Ojeda 15 Ojeda 30 Ojeda 164 Ojeda 35  UBCBG, No. 034350-0077-1998 FTBG, No. 2001-0108B  Ojeda 133 Ojeda 110  UBCBG, No. 032961-0614-1996 UBCBG, No. 036707-0640-2003  Ojeda 59 Ojeda 63  Voucher specimen, UBC herbarium  Clarke 16-01  MIMOSOIDEAE Mimosoid crown node Tribe Acacieae Acacia tortuosa (L.) Willd. Tribe Ingeae Albizia julibrisii Durazz. Calliandra haematocephala Hassk. Inga paterno Harms Lysiloma sabicu Benth. Pithecellobium arboreum (L.) Urb. Tribe Mimoseae Leucaena leucocephala (Lam.) de Wit  Genistioid crown node Tribe Crotalarieae Crotalaria pallida Aiton Crotalaria sp. Tribe Genisteae Adenocarpus decorticans Boiss. *Chamaecytisus albus Rothm. Chamaespartium sagitale (L.) P. Gibbs Cytisus nigricans L. C. scoparius (L.) Link Erinacea anthyllis Link Genista lydia Boiss. G. pilosa L. G. radiata DC. G. tenera (Jacq.) Kuntze Laburnum x watereri cultivar Vossii Lupinus littoralis Douglas ex Lindl. L. polyphyllus C.E. Anderson *Retama rhodorhizoides Webb & Berthel. Spartium junceum L. Ulex europaeus L. Tribe Sophoreae Sophora davidii (Franch.) Skeels Sophora tomentosa L. Tribe Thermopsideae Baptisia australis (L.) R. Br. Thermopsis macrophylla Hook. & Arn. Dalbergioid crown node Tribe Adesmieae +Adesmia atacamensis Phil.  59  Voucher information  Taxon  Collection site and number  Tribe Amorpheae +Amorpha georgiana Wilbur  Voucher specimen, UBC herbarium  A. herbacea Walter +Apoplanesia paniculata C. Presl +Dalea carthagenensis (Jacq.) J.F. Macbr.  MBG, No. 2005-0070 Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium  +D. eysenhardtioides Hemsl. +D. greggii A. Gray +D. leporina (Aiton) Bullock +Marina diffusa (Moric.) Barneby +M. ghiesbreghtii Barneby  Voucher specimen, CICY herbarium Voucher specimen, UBC herbarium Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium  +Psorothamnus arborescens (Torr. ex A. Gray) Barneby Tribe Dalbergieae +Andira galeottiana Standl.  Voucher specimen, UBC herbarium  +A. inermis (W. Wright) Kunth ex DC. +Aeschynomene americana L. +A. ciliata Vogel  Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium  +A. fascicularis Schltdl. & Cham. *Arachis hypogaea L. Brya ebenus (L.) DC. +Dalbergia ecastaphyllum (L.) Taub. +D. brownei (Jacq.) Schinz +D. glabra (Mill.) Standl.  Voucher specimen, CICY herbarium Propagated from comercial seeds FTBG, No. 992154A Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium  +D. glomerata Hemsl. +D. stevensonii (Standl.) +D. violacea (Jacq.) Hoffmanns. *Diphysa carthagenensis Benth. ex Benth. & Oerst. +Pterocarpus acapulcensis Rose +P. hayesii Hemsl. Stylosanthes hamata (L.) Taub. *Tipuana tipu (Benth.) Kuntze  Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium Voucher specimen, UBC herbarium JBRCICY  Gutierrez B.C. 7740 Tenorio P. 19415 A. Mendez t. 6786 Magaña M.H. & Correl C. 613 E. Cabrera 15268 Ojeda 125 I. Espejel 476 C. Chan 6484 J. Flores & E. Ucan E. 8303 G. Aguilar M. 413 Gutierrz B.C. 6675 Mimura 227 Ojeda 143  Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium FTBG, without number JAO, No. 0680-70  R. Torres 11961 C:P. Cowan 4654 Ojeda 118 -  Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium  Beamish 1314 Beamish 418 Beamish 675  Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium Voucher specimen, UBC herbarium  Dempster 3075 Beamish 155 Beamish 239 Beamish 674 Dempster 3427  Wild collected, Coral Gables, Florida UBCBG, No. 011143-0013-2002  Ojeda 99 Ojeda 136  UBCBG, No. 007973-0241-1974  Ojeda 137  Mirbelioid crown node Tribe Mirbelieae +Daviesia ulicifolia Andrews +Gastrolobium floribundum S. Moore +G. subcordatum (Benth.) G. Chandler & Crisp +Gompholobium grandiflorum Andrews +G. virgatum Sieber ex DC. +Isotropis cuneifolia Benth. ex B.D. Jacks. +Mirbelia dilatata R. Br. +Oxylobium scandens (Small) Benth. Milletioid crown node Tribe Desmodieae Desmodium incanum DC. Lespedeza thunbergii (DC.) Nakai Tribe Indigofereae Indigofera decora Lindl.  Voucher specimen, CICY herbarium  Leonard and Moore 1720 Ojeda 94 J. S. Flores 9885 Ventura E. & Lopez E. 4138 D. Tenorio 2114 Helmkamp 5619 J. Martinez 1021 J.I. Calzada 20413 Mendez Tun A. 7159 Sanders 2707  60  Taxon  Collection site and number  I. kirilowii Maxim. ex Palibin Tribe Phaseoleae *Canavalia rosea (Sw.) DC. Centrosema virginianum (L.) Benth. Clitoria fairchildiana R.A. Howard C. ternatea L. *Erythrina carnea Blanco *E. crista-galli L. *E. indica Lam. *Erythrina sp. *E.corallodendron L. Galactia striata (Jacq.) Urb. Glycine max (L.) Merr. *Vigna elegans (Piper) Maréchal, Mascherpa & Stainier V. luteola (Jacq.) Benth. Phaseolus coccineus L. var. Scarlet Runner P. vulgaris L. var. Blue Lake Tribe Psoraleae *Bituminaria bituminosa (L.) C.H. Stirt. Tribe Millettieae *Piscidia piscipula (L.) Sarg. Robinioid crown node Tribe Loteae Anthyllis hermanniae L. Coronilla valentina L. C. varia L. Hosackia chihuahuana S. Watson Lotus arenarius Brot. L. burttii Borsos  UBCBG, No. 33916-598-98  Voucher information Ojeda 77  Wild collected, Yucatán, México MBG, No. 2003-1125 MBG, No. 84 MBG, No. 2005-1338ª JAO, No. 254-99 JAO, No. 0222-94 JBRCICY JAO, No. 504-99 Cultivated at La Palma, Spain Wild collected, Big Pine Key, Florida Propagated from comercial seeds Wild collected, Dzitya, Yucatán, México  RD 2229 Ojeda 97 Ojeda 84 Ojeda 91 Ojeda 162 Ojeda 163 Ojeda 140 Ojeda 165 Ojeda 157 Ojeda 120 Ojeda 76 Ojeda 144  MBG, without number Propagated from comercial seeds Propagated from comercial seeds  Ojeda 89 Ojeda 36 Ojeda 38  Wild collected, Gran Canaria, Spain  Ojeda 166  JBRCICY  Ojeda 139  UBCBG, No. 035419-0389-2000 UBCBG, without number UBC campus, Vancouver From seeds, Pullman No. 18085 From seeds, Pullman KBG 5688 Cultivated from seeds, Miyasaki University UBC campus, Vancouver From seeds, Pullman KBG 5810 Propagated from seeds UBCBG, No. 032962-0447-1996 Propagated from seeds Propagated from seeds Propagated from seeds  Ojeda 138 Ojeda 33 Ojeda 39 Ojeda 79 Ojeda 78 Ojeda 72  Wild collected, Dzitya, Yucatán, México JBRCICY Voucher specimen, CICY herbarium UBC campus, Vancouver UBCBG, without number  Ojeda 146 Ojeda 141 G. Aguilar 629 Ojeda 75 Ojeda 63  Voucher specimen, CICY herbarium Voucher specimen, CICY herbarium FTBG, No. 2005-0522A  M. Narváez 1400 Duran et al 3506 0jeda 124  Propagated from seeds  Ojeda 73  Wild collected, Spanish Bank beach, Vancouver Wild collected UBC campus, Vancouver Propagated from comercial seeds Propagated from comercial seeds Wild collected, UBC campus, Vancouver  Ojeda 16  L. corniculatus L. L. eriosolen (Maire) Mader and Podlech L. filicaulis Durieu L. hirsutus L. L. japonicus (Regel) K. Larsen Gifu B-129 *L. japonicus (Regel) K. Larsen Gifu B-129 L. japonicus (Regel) K. Larsen MG 20 Tribe Robineae *Coursetia caribea (Jacq.) Lavin *Gliricidia maculata (Kunth) Walp. +G. sepium (Jacq.) Kunth ex Walp. Robinia hispida L. R. pseudoacacia L. Tribe Sesbanieae *Sesbania grandiflora (L.) Pers. *S. herbaceae (Mill.) McVaugh S. punicea (Cav.) Benth. IRLC crown node Tribe Cicereae Cicer arietinum L. Tribe Fabeae Lathyrus japonicus Willd. L. latifolius Visiani L. odoratus L. cultivar April in Paris L. sativus L. cultivar Electric Blue L. sylvestris L.  Ojeda 27 Ojeda 243 Ojeda 71 Ojeda 58 Ojeda 70 Ojeda 70 Ojeda 69  0jeda 31 Ojeda 43 Ojeda 80 Ojeda 29  61  Taxon  Collection site and number  L. venetus Rouy L. vernus (L.) Bernh. Lens culinaris Medik. Pisum sativum L. Vicia cracca L. V. faba L. V. hirsuta (L.) Gray V. nigricans Hook. & Arn. V. sativa L. Medicago sativa L. M. lupina L. Melilotus albus Medik. M. officinalis (L.) Pall. Ononis spinosa L. Trifolium dubium Sibth. T. hybridum L. T. incarnatum L. T. repens L. T. wormskioldii Lehm. Tribe Galegeae Astragalus crassicarpus Nutt. Tribe Millettieae Wisteria brachybrotis Siebold & Zucc. W. floribunda (Willd.) DC. var .violacea W. floribunda (Willd.) DC. var. rosea W. sinensis (Sims) Sweet  UBCBG, No. 036759-0684-2003 UBCBG, No. 032718-0620Propagated from comercial seeds Propagated from comercial seeds Wild collected, UBC campus, Vancouver Propagated from comercial seeds Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver UBCBG, No. 034702-0433-1999 Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver UBCBG, without number Wild collected, UBC campus, Vancouver Wild collected, UBC campus, Vancouver  Voucher information Ojeda 60 Ojeda 52 Ojeda 74 Ojeda 44 Ojeda 18 Ojeda 37 Ojeda 28 Ojeda 24 Ojeda 26 Ojeda 34 Ojeda 25 Ojeda 17 Ojeda 19 Ojeda 132 Ojeda 22 Ojeda 23 Ojeda 32 Ojeda 21 Ojeda 16  UBCBG, No. 037261-0653-2004  Ojeda 135  UBC campus, Vancouver UBCBG, No. 022727-0481-1983 UBCBG, No. 037625-0075-2005 UBCBG, No. 014050-0223-11976  Ojeda 68 Ojeda 66 Ojeda 67 Ojeda 65  62  Table 3.2 Distribution of the major epidermal types in each of the three types of petal in the Leguminosae. PCS= papillose conical cells, PKR= papillose knobby with rugose sculpture, PLS = papillose lobular cells, TRG= tabular rugose with a granular scupture, TRS= tabular rugose striate, (TRSi TRSii TRSiii represents minor variations within this epidermal type that allow petal identification within the same species) and TFS= tabular flat longitudinally striate. More than one major epidermal type with more or less equal distribution are separated by a slash. Cell types in bold indicates the side more differentiated in each type of petal. If both sides are bold, then more or less the same level of differentiation is implied. PLS is a rare type that is never characteristic of whole petals and its presence is therefore only noted by the symbol ‡ (it is found in some Lathyrus species where it is restricted to the margin of the dorsal petal). s= stomata, t= trichomes, st= trichomes and stomata. *= flowers preserved in ethanol and analyzed using a ligth microscope, + = flowers re-dehydrated from voucher specimens, preserved in ethanol and analyzed using a light microscope. Other species were studied using the SEM and fresh material. -= petals absent.  63  Taxon POLYGALACEAE Polygala chamaebuxus P. diversifolia P. grandiflora P. myrtifolia  LEGUMINOSAE CAESALPINOIDEAE Cercis crown node Tribe Cercidae Cercis canadensis C. yunnanensis Bauhinia divaricata B. natalensis B. petersiana B. tarapotensis B. tomentosa B. variegata Tylosema fassoglensis Caesalpioid crown node Tribe Detarieae * Brownea ariza B. capitella Tamarindus indica Umtizia crown node Tribe Caesalpineae *Gleditsia triacanthos Tribe Cassieae Cassia emarginata *C. grandis C. fistula C. javanica *C. javanica C. nodosa C. roxburghii Chamaecrista lineata Senna alata S. corymbosa S. ligustrina S. mexicana *S. mucronifera *S. nicaraguensis *S. pallida S. polyphylla *S. racemosa S. septentrionalis *S. siamea *S. wislizeni Tribe Caesalpineae Caesalpinia pulcherrima *C. gaumeri Delonix regia Peltophorum pterocarpum Tara cacalao MIMOSOIDEAE Mimosoid crown node Tribe Acacieae Acacia tortuosa  Petaloid sepal abaxial adaxial TRS TRSt TRS TRS TRSs TRSs PKR PKR  Dorsal petal abaxial adaxial PCS PCS TRS TRS TRS TRS TFS TFS  Ventral petal abaxial adaxial PCS PCS TFS TFS TRS TRS TFSt TFSt  Dorsal petal abaxial adaxial  Lateral petal abaxial adaxial  Ventral petal abaxial adaxial  TRS TRS TRS TRS TRS PKR PCS PCS PCSt  TRS TRS TRS TRS TRSs PKR TRS PCS PCSt  TRS TRS TRS TRS TRS PKR PCS PCS PCSt  TRS TRS TRS TRS TRSs PKR TRS PCS PCSt  TRS TRS TRS TRS TRS PKR PCS PCS PCSt  TRS TRS TRS TRS TRSs PKR TRS PCS PCSt  PKR TRS PKR  PKR TRS PKR  PKR TRS PKR  PKR TRS PKR  PKR TRS PKRt  PKR TRS PKRt  TRSt  TRSt  TRSt  TRSt  TRSt  TRSt  TRSst PKRt PKRst PKRt PKRt PKR PKRt TRSst PCS PKR TRS PKR TRSt TRS TRSt TRS TRS TRS TRS TRS  TRS PKR PKR PKRt PKRt PKRs PKR TRSt PKR PKR TRS PKR TRSt TRS TRSt TRS TRS TRS TRS TRSst  TRSst PKR PKRst PKRt PKRt PKR PKR TRSst PCS PKR TRS PKR TRSt TRS TRSst TRS TRS TRS TRS TRS  TRS PKR PKR PKRt PKRt PKR PKR TRSt PKR PKR TRS PKR TRSt TRS TRSst TRS TRS TRS TRS TRS  TRSs PKR PKRst PKRt PKRt PKR PKR TRSt PCS PKR TRS PKR TRSt TRS TRSst TRS TRS TRS TRS TRS  TRS PKR PKR PKRt PKRt PKR PKR TRSt PKR PKR TRS PKR TRSt TRS TRSst TRS TRS TRS TRS TRS  PCS TRS TRSt TRS TRS  PCS TRS TRS TRS TRSt  PCS TRS TRS TRS TRS  PCS TRS TRS TRS TRS  PCS TRS TRS TRS TRSt  PCS TRS TRS TRS TRS  TRG  TRG  TRG  TRG  TRG  TRG  64  Taxon Tribe Ingeae Albizia julibrissin Calliandra haematocephala Inga paterno Lysiloma sabicu Pithecellobium arboreum Tribe Mimoseae Leucaena leucocephala PAPILIONOIDEAE Swartzia crown node Tribe Swartzieae +Swartzia pittieri Cladastris crown node [Tribe Sophoreae] Cladrastis lutea C. sinensis Genistioid crown node Tribe Crotalarieae Crotalaria pallida Crotalaria sp. Tribe Genisteae Adenocarpus decorticans *Chamaecytisus albus Chamaespartium sagittale Cytisus nigricans Cytisus scoparius Erinacea anthyllis Genista lydia G. pilosa G. radiata G. tenera Laburnum x watereri Lupinus littoralis L. polyphyllus *Retama rhodorhizoides Spartium junceum Ulex europaeus Tribe Sophoreae Sophora davidii S. tomentosa Tribe Thermopsideae Baptisia australis Thermopsis macrophylla Dalbergioid crown node Tribe Adesmieae +Adesmia atacamensis Tribe Amorpheae +Amorpha georgiana +A. herbacea +Apoplanesia paniculata +Dalea carthagenensis +D. eysenhardtioides +D. greggii +D. leporina +Marina diffusa +M. ghiesbreghtii +Psorothamnus arborescens  Dorsal petal abaxial adaxial  Lateral petal abaxial adaxial  Ventral petal abaxial adaxial  TRGt TRG TRGt TRGst TRG  TRG TRG TRGst TRGst TRG  TRGt TRG TRGt TRGst TRG  TRG TRG TRGst TRGst TRG  TRGt TRG TRGt TRGst TRG  TRG TRG TRGst TRGst TRG  TRG  TRGst  TRG  TRGst  TRG  TRGst  PKR  PKR  -  -  -  -  PKR PKR  PKR PKR  PKR PKR  PKR PKR  TFS TFS  TFS TFS  TRS TRS  TRS TRS  PKR PKR  TRS PKR  TFS TRS  TFS TRS  TRS TRS TRS PCS PKR PKR TRS PKR TRSt PKR PCS PCS PKR TRS PKR TRS  TRS TRS TRS TRS TRS PKR TRS TRSt PKR PKR PCS PCS PKR TRS PKR PKR  TRS TRS PCS PCS PCS PKR PCS PCS PCS PCS PCS TRS PKR PCS PCSt PCS  TRS TRS PCS PCS PCS PKR PCS PCS PCS PCS PCS TRS PKR TRS PCS PCS  TFS TRSt/TFS TFSt TFS TFS TFS TFS TFSt TFSt TRS TFS TRS/TFS TRS/TFS TRSt TFSt TFS  TFS TRSt TFS TFS TFS TFS TFS TFS TFS TFS TFS TRS/TFS TRS/TFS TRSt TFS TFSt  PCS PKRs  PCS PKR  PCS PKR  PCS TRSs  TFSs TFS  TFS TFSs  PKR PKR  TRS TRS  PKR PCS  TRSs PKR  TFS TFS  TFS TFS  PCS  PCS  PCS  PCS  TFS  TFS  TRS TRS TRSi TRSi TRSi TRSi PCS TRSi TRSi TRSi  TRS TRS TRSi TRSi TRSi TRSi PCS TRSi TRSi TRSi  TRSi TRSi TRSi TRSi PCS TRSi TRSi TRSi  TRSi TRSi TRSi TRSi PCS TRSi TRSi TRSi  TRSi TRSi TRSi TRSi TRSi TRSi TRSi TRSi  TRSi TRSi TRSi TRSi TRSi TRSi TRSi TRSi  65  Taxon Tribe Dalbergieae +Andira galeottiana +A. inermis +Aeschynomene americana +A. ciliata +A. fascicularis *Arachis hypogaea Brya ebenus +Dalbergia ecastaphyllum +D. brownei +D. glabra +D. glomerata +D. stevensonii +D. violacea *Diphysa carthaginensis +Pterocarpus acapulcensis +P. hayesii Stylosanthes hamata *Tipuana tipu Mirbelioid Tribe Mirbelieae +Daviesia ulicifolia +Gastrolobium floribundum +G. subcordatum +Gompholobium grandiflorum +G. virgatum +Isotropis cuneifolia +Mirbelia dilatata +Oxylobium scandens Milletioid crown node Tribe Desmodieae Desmodium incanum Lespedeza thunbergii Tribe Indigofereae Indigofera decora I. kirilowii Tribe Phaseoleae *Canavalia rosea Centrosema virginianum Clitoria fairchildiana C. ternatea *Erythrina carnea *E. crista-galli *E. indica *Erythrina sp. *E. corallodendron Galactia striata Glycine max *Vigna elegans V. luteola Phaseolus coccineus P. vulgaris Tribe Psoraleeae *Bituminaria bituminosa  Dorsal petal abaxial adaxial  Lateral petal abaxial adaxial  Ventral petal abaxial adaxial  PKR PCS PCS PCS TRS TRS PKR PCS PKR PCS PCS PCS PKR PKR PCS PCS TRS TRS  PCS PCS PCS PCS TRSt TRS PKR PCS PKR PCS PCS PCS PKR PKR PCS PCS TRS TRS  PCS PCS PCS PCS PCS TRS PKR PCS PCS PCS PCS PCS PKR TRS TRS PCS PCS PKR  PCS TRS PCS PCS PCS TRS PKR PCS TRS PCS PCS PCS PKR TRS TRS PCS TRS PKR  TRS TRS TRS/TFSt TRS TFS TFS TRS TRS TRS TRS TRS TRS TFS TFS TRS TRS TFS TRS  TRS TRS TRS/TFS TFS TFS TFS TRS TRS TRS/TFS TRS TRS TRS TFS TFS TRS TRS TFS TRS  PKR PKR PKR PKR  PKR PKR PKR PKR  PKR PKR PKR PCS  PKR PKR PKR PCS  TFS TRS TRS TRS  TFS TRS TRS TRS  PKR PKR PKR PKR  PKR PKR PKR PKR  PKR PKR PKR TRS  PKR PKR PKR PKR  TRS TFS TRS TRS  TRS TFS TRS TRS  PCS PCS  PCS PCS  PCS PCS  TRS PCS  TRS TFS  TRS TFS  TRSi TRSi  TRSi TRSi  TRSi TRSi  TRSi TRSi  TRSi TRSi  TRSi TRSi  PKR PCSt TRSt TRS PCS PCSs PCS PCS PCS PKR PKR TRSs PKRs PCS PCS  PKR PCS PCSt PCS PCS PCS PCS PCS PKR PKR PKR PKR PKR PCS PCS  TRS PCSt TRSt PKR PKR PCS PCS PKR PKR PKR PCS TRS PKR PCS PCS  TRS PCS TRSt TRS PKR PCS PKR PKR PKR PKR PCS TRS PKR PCS PCS  TRS TFSt TFSt TFS PKR PCSs PKR PKR PKR TRS TRS TFS TFS TFS TFSt  TRS TRS/TFS TFS TFS PKR PCSs PKR PKR PCS TRS TRS TFS TFS TFS TFS  PCS  PCS  PKR  PCS  TRS  TRS  66  Taxon Tribe Millettieae *Piscidia piscipula Robinoid crown node Tribe Loteae Anthyllis hermanniae Coronilla valentina C. varia Hosackia chihuahuana Lotus arenarius L. burttii L. corniculatus L. eriosolen L. filicaulis L. hirsutus L. japonicus Gifu B-129 *L. japonicus Gifu B-129 L. japonicus MG 20 Tribe Robineae *Coursetia caribaea *Gliricidia maculata +G. sepium Robinia hispida R. pseudoacacia Tribe Sesbanieae *Sesbania grandiflora *S. herbacea S. punicea IRLC crown node Tribe Cicereae Cicer arietinum Tribe Fabeae Lathyrus japonicus L. latifolius L. odoratus L. sativus L. sylvestris L. venetus L. vernus Lens culinaris Medicago sativa M. lupina Melilotus albus M. officinalis Ononis spinosa Pisum sativum Trifolium dubium T. hybridum T. incarnatum T. repens T. wormskioldii Vicia cracca V. faba V. hirsuta V. nigricans V. sativa Tribe Galegeae Astragalus crassicarpus Tribe Millettieae Wisteria brachybrotis  Dorsal petal abaxial adaxial  Lateral petal abaxial adaxial  Ventral petal abaxial adaxial  PKRt  PKR  PKR  PKR  TFS  TFS  PCS PCS PCS PCS PCS PCS PCS TRS PCS PCS PCS PCS PCS  PCS PCS PCS PCS PCS PCS PCS PCS PCS PCS PCS PCS PCS  TRS PCS PCS PCS TRS TRS TRS TRS TRS TRS TRS TRS TRS  TRS PCS TRS TRS TRS TRS TRS TRS TRS TRS TRS TRS TRS  TFS TRS/TFS TFS TRS/TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS  TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS  PKR PKR PKR PKR PKR  PKR PKR PKR PKR PKR  PKR PKR PKR PKR PKR  TRS PKR PKR PKR PKR  TRS TFS TFS TFS TFS  TFS TFS TFS TFS TFS  TRS PCS PKR  TRS PCS PKR  PKR TRS TRS/PKR  TRS TRS TRS/PKR  TFS TFS TFS  TFS TFS TFS  TRSi  TRSi  TRSi  TRSi  TRSiii  TRSiii  TRSi TRSi TRSi TRSi TRSi PKR TRSi TRSi TRS TRSi TRSi TRSi TRSt TRSi PKR TRSi TRSi TRSi TRSi TRSi TRSi TRSi TRSi TRSi  TRSi TRSi TRSi TRSi TRSi ‡ PKR TRSi ‡ TRSi PCS TRSi TRSi TRSi TRS TRSi PCS TRSi TRSi TRSi TRSi TRSi TRSi TRSi PCS PCS  TRSi TRSi TRSi/PKR TRSi TRSii PKR PKR TRSi PCS TRSi TRSi TRSi TRS TRSii TRSi TRSi TRSi TRSi TRSi PCS PCS TRSi PCS PCS  TRSi TRSi TRSi TRSi TRSii PKR PKR TRSi PCS TRSi TRSi TRSi TRS TRSii TRSi TRSi TRSi TRSi TRSi TRSii PCS TRSi TRSi TRSi  TRSiii TRSiii TRSiii TRSiii TRSiii TRSiii TRSiii TRSi TRS TRSi TRSi TRSi TFS TRSiii TRSit TRSi TRSi TRSiii TRSiii TRSi TRSiii TRSiii TRSi TRSi  TRSiii TRSiii TRSiii TRSiii TRSiii TRSiii TRSiii TRSi TRS TRSi TRSi TRSi TFS TRSiii TRSi TRSi TRSi TRSiii TRSiii TRSi TRSiii TRSiii TRSi TRSi  TRSi  TRSi  TRSi  TRSi  TRSi  TRSi  PKR  PKR  PKR  TRSii  TFS  TFS  67  Taxon W. floribunda var . violacea W. floribunda var. rosea W. sinensis  Dorsal petal abaxial adaxial PKR PKR PKR PKR PKR PKR  Lateral petal abaxial adaxial PKR PKR PKR PKR TRSii PKR  Ventral petal abaxial adaxial TFS TFS TFS TFS TFS TFS  Table 3.3 Classification of the epidermal types observed in Leguminosae. Major type group  Papillose  Tabular  Major epidermal type  Abbreviation  Figure  Example  Papillose conical cells with PCS 3 D, J, P Lotus japonicus striations (standard) Papillose knobby cells with a PKR 3 E, K, Q Robinia pseudoacacia rugose sculpture (standard and lateral) Papillose lobular cells with PLS 3 F, L, R Lathyrus venetus striations (standard) Tabular rugose cells with TRS 3 B, H, N Wisteria sinensis longitudinal striations (wings) Tabular rugose cells with a TRG 3 A, G, M Calliandra haematocephala granulose sculpture (all petals) Tabular flat cells with TFS 3 C, I, O Lotus japonicus longitudinal striations (keel) Note: TRS is the most variable type. The main subtypes may be distinguished as follows: (i) cells elongated with dense striation (ii) cells more or less isodiametric with dense striation and (iii) isodiametric or elongated cells with less dense striations  Table 3.4 Distribution of the major epidermal cell types within sampled Papilionoideae. PCS= papillose conical cells, PKR= papillose knobby cells, TRS= tabular rugose cells with striations, TFS= tabular flat cells with striations.  Petal Dorsal Lateral Ventral  PCS Abaxial Adaxial 38 45 44 33 1 2  PKR Abaxial Adaxial 42 40 38 32 4 3  TRS Abaxial Adaxial 53 48 48 66 61 61  TFS Abaxial Adaxial 0 0 0 0 63 64  68  Figure 3.1 Diversity of flower symmetry in the Leguminosae. (A) Zygomorphic flowers of Cercis canadensis (Caesalpinioideae). (B) Flowers of Bauhinia tomentosa (Caesalpinioideae). (C) Radially symmetric flower of Brownea capitella (Caesalpinioideae). (D) Zygomorphic flowers of Tara cacalao (Caesalpinioideae). (E) Asymmetric enantiostylous flower in Cassia emarginata (Caesalpinioideae). (F) Zygomorphic flower with the dorsal petal differentiated with respect to the other petals in Caesalpinia pulcherrima (Caesalpinioideae). (G) Radial flower with reduced petals in Inga paterno (Mimosoideae). (H) Zygomorphic flowers in Lespedeza thunbergii (Papilionoideae) (with the ventral petal more exposed), (I) Clitoria ternatea (Papilionoideae) (the dorsal petal is pointing downwards), (J) Sesbania punicea (Papilionoideae), and (K) Lathyrus sylvestris (Papilionoideae) (ventral petals enclosed).  69  Figure 3.2 Four axes of variation considered in the study of the epidermal types and their distribution on each petal. A) Lotus corniculatus, three types of petals in zygomorphic papilionoid flowers, a) dorsal, b) lateral and c) ventral, within an adaxial-abaxial axis within the flower. B) Abaxial-adaxial surface within the petal. The abaxial side is exposed in lateral (b) and ventral (c) petals in most papilionoids. However, the adaxial surface is exposed in most standard petals (a). C) Senna corymbosa with three types of petals, (a) dorsal, (b) lateral and c) ventral within an adaxial-abaxial axis within the flower. Further axes are: proximal-distal and medio-lateral axes within D) the dorsal petal, E) lateral petal and F) ventral petal in Lotus corniculatus. The base of the dorsal petal (claw) in L. corniculatus has been separated from the rest of the petal.  70  Figure 3.3 Classification of the major epidermal types in Leguminosae. (A, G, M) tabular rugose cells with granulose sculpture (TRG) in Calliandra haematocephala (Mimosoideae). (B and H) tabular rugose cells with striation (TRS) in Wisteria sinensis (Papilionoideae) and (N) in Lotus japonicus (Papilionoideae). (C and I) tabular flat cells with striations (TFS) in W. sinensis and (O) in Lotus japonicus. (D, J, P) papillose conical cells (PCS) in the dorsal petal of Lotus japonicus. (E, K, Q) papillose knobby cells (PKR) in the dorsal and lateral petals of Robinia pseudoacacia (Papilionoideae). (F and L) papillose lobular cells (PLS) in the dorsal petal of Lathyrus venetus and (R) in Lathyrus sylvestris (Papilionoideae). This latter epidermal type was only observed in these two species. All images correspond to the adaxial side of the petal. Scale bar 50 µm (A-F), 100 µm (G-L) and 20 µm (M-R).  71  Figure 3.4 Minor epidermal types within the tabular rugose cells with striations (TRS). Minor epidermal variants within TRS are designated as i, ii and iii. (A-C) variation among minor epidermal types in Pisum sativum (Papilionoideae) enables each petal type to be distinguished. Variation within minor epidermal types distinguished the ventral but does not distinguish between the lateral and dorsal petals in (D-F) Vicia hirsuta (Papilionoideae), (G-I) Trifolium repens (Papilionoideae). Variation among minor epidermal types does not allow clear characterization of the three petal types in (J-L) Melilotus officinalis (Papilionoideae) and (M-O) Cercis canadensis (Caesalpinioideae). All illustrations correspond to the adaxial side of the petal, except H, I, K, L, N and O, which correspond to the abaxial side. All scale bars 50 µm.  72  Figure 3.5 Distribution of the epidermal types along the dorsiventral (adaxial-abaxial) axis within the flower. A) micromorphological variation in Lotus burttii (as in almost all Loteae) (Papilionoideae), B) lack of micromorphological variation of major epidermal types in Cassia roxburghii (Caesalpinioideae) with only PKR, C) Cassia emarginata (Caesalpinioideae) with only TRS, and D) Senna alata (Caesalpinioideae) with only PCS on all petals. All petals have the adaxial side shown, except L. burttii, where the abaxial side is presented on lateral and ventral petals and Senna alata where images shown the abaxial side. Scale bars A) 20 µm, B) 50 µm, C) 100 µm, D) 50 µm.  73  Figure 3.6 Schematic representation of the phylogenetic relationships within Leguminosae showing the typical distribution of the major epidermal types observed. This figure is intended to summarise the main patterns but it should be noted that rare variant patterns may occur in clades as well as those listed. The general epidermal surface observed within the flower is given in relation to the dorsiventral axis within the flower, using the representation: dorsal/lateral/ventral petal. Clades with an asterisk contain lineages with loss of papillose cells (PCS and PKR). The numbers following the asterisk indicate the number of losses under parsimony and ML, respectively. -- lack of this type of petal (Tree according to Wojciechowski et al., 2004; Lavin et al., 2005; Lewis et al., 2005). IRLC=inverted repeat-lacking clade.  74  Figure 3.7 Distribution of epidermal types along a proximo-distal axis within the same petal. (A-E) Lupinus littoralis with transitions from poorly differentiated cells at the base of the adaxial side in the dorsal petal (C), to papillose cells (PCS) on the central part and apex of the petal (A,B). (F-J) the abaxial side of the dorsal petal in Lotus japonicus with transitions from poorly differentiated cells at the base of the petal (H) to PCS in the central and distal regions (G, H). A photograph of the standard petal of Lotus japonicus is shown besides the images with the claw separated and shown below. (K-P) the adaxial side of the dorsal petal in Lathyrus sylvestris has a transition zone from TRS to PLS on the borders of the petal with a transition zone where cells have a mixture of morphological features of both epidermal types. (Q-V) the abaxial side of the lateral petal in Lotus japonicus where TRS is mainly observed at the base and in the central part of the petal and there is a transition zone from TRS to PCS where the cells have a mixture of both epidermal types. Scale bars 500 µm in B; 200 µm in G; 100 µm in A, C, F, J-M; 50 µm in D, H, I, O, P, Q-S; 20 µm in E, N, T-V.  75  76  3.5 Bibliography Avila, A., Nieto, C., Cañas, L., Benito, J., Paz-Ares, J., 1993. Petunia hybrida genes related to the maize regulatory C1 gene and to the animal MYB proto-oncogenes. Plant Journal 3, 553-562. Baagøe, J., 1977. Taxonomical applications of ligule microcharacters in Compositae. 1. Anthemideae, Heliantheae and Tageteae. Botanisk Tidsskrift 71, 192-223. Baagøe, J., 1980. SEM-studies in ligules of Lactuceae (Compositae). Botanisk Tidsskrift 75, 199-217. Barthlott, W., 1981. Epidermal and seed surface characters of plants: systematic applicability and some evolutionary aspects. Nordic Journal of Botany 1, 345-355. Barthlott, W., 1990. Scanning electron microscopy of the epidermal surface in plants. In: Claugher, D. (Ed.), Scanning electron microscopy in taxonomy and functional morphology. Clarendon Press, Oxford, pp. 69-83. Barthlott, W., Ehler, N., 1977. Raster-Elektronemmikroskopie der Epidermis-Oberflächen von Spermatophyten. Tropische Subtropische Pflanzenwelt 19, 367-467. Choi, B., Seok, D., Endo, H.O., 2006. Phylogenetic significance of stylar features in genus Vicia (Leguminosae): an analysis with molecular phylogeny. Journal of Plant Research 119, 513523. Christensen, K., Hansen, H., 1998. SEM-studies of epidermal patterns of petals in the angiosperms. Opera Botanica 135, 1-91. Citerne, H.L., Luo, D., Pennington, R.T., Coen, E., Cronk, Q.C.B., 2003. A phylogenomic investigation of CYCLOIDEA-like TCP genes in the Leguminosae. Plant Physiology 131, 1042-1053.  77  Citerne, H.L., Pennington, R.T., Cronk, Q.C., 2006. An apparent reversal in floral symmetry in the legume Cadia is a homeotic transformation. Proceedings of the National Academy of Sciences United States of America 103, 12017-12020. Comba, L., Corbet, S., Hunt, H., Outram, S., Parker, J.S., Glover, B.J., 2000. The roles of genes influencing the corolla in pollination of Antirrhinum majus. Plant, Cell and Environment 23, 639-647. Cronk, Q.C., 2006. Legume flowers bear fruit. Proceedings of the Natural Academy of Sciences of the United States of America 103, 4801-4802. Degtjareva, G.V., Kramina, T.E., Sokoloff, D.D., Samigullin, T.H., Valiejo-Roman, C.M., Antonov, A.S., 2006. Phylogeny of the genus Lotus (Leguminosae, Loteae): evidence from nrITS sequences and morphology. Canadian Journal of Botany 84, 813-830. Dyer, G.A., Whitney, H.M., Arnold, S.E.J., Glover, B.J., Chittka, L., 2006. Behavioural ecology: Bees associate warmth with floral colour. Nature 442, 525. Dyer, G.A., Whitney, H.M., Arnold, S.E.J., Glover, B.J., Chittka, L., 2007. Mutations perturbing petal cell shape and anthocyanin synthesis influence bumblebee perception of Antirrhinum majus flower colour. Arthropod-Plant Interactions 1, 45-55. Ellison, N.W., Liston, A., Steiner, J.J., Williams, W.M., Taylor, N.L., 2006. Molecular phylogenetics of the clover genus (Trifolium Leguminosae). Molecular Phylogenetics and Evolution 39, 688-705. Endress, P.K., 2001. Evolution of floral symmetry. Current Opinion in Plant Biology 4, 86-91. Feng, X., Zhao, Z., Tian, Z., Xu, S., Luo, Y., Ca, Z., Wang, Y., Yang, J., Wang, Z., Weng, L., Chen, J., Zheng, J., Guo, X., Luo, J., Sato, S., Tabata, S., Ma, W., Cao, X., Hu, X., Sun, C., Luo, D., 2006. Control of petal shape and flower zygomorphy in Lotus japonicus.  78  Proceedings of the National Academy of Sciences of the United States of America 103, 4970-4975. Fu, Y., Gu, Y., Zheng, Z., Wasteneys, G., Yang, Z., 2005. Arabidopsis interdigitating cell growth requires two antagonistic pathways with opposing action on cell morphogenesis. Cell 120, 687-700. Glover, B.J., Martin, C., 1998. The role of petal shape cell and pigmentation in pollination success in Antirrhinum majus. Heredity 80, 778-784. Glover, B.J., Martin, C., 2002. Evolution of adaptive petal cell morphology. In: Cronk, Q.C.B., Bateman, R.M., Hawkins, J.A. (Eds.), Developmental genetics and plant evolution. Taylor & Francis, London, pp. 160-172. Glover, B.J., Perez-Rodriguez, M., Martin, C., 1998. Development of several epidermal cell types can be specified by the same MYB-related plant transcription factor. Development 125, 3497-3508. Gorton, H.L., Vogelmann, T.C., 1996. Effects of epidermal cell shape and pigmentation on optical properties of Antirrhinum petals at visible and ultraviolet wavelengths. Plant Physiology 112, 879-888. Guimil, S., Dunand, C., 2007. Cell growth and differentiation in Arabidopsis epidermal cells. Journal of Experimental Botany 58, 3829-3840. Hammett, K.R.W., Muray, B.G., Markham, K.R., Hallet, I.C., 1994. Interspecific hybridization between Lathyrus odoratus and Lathyrus belinensis. International Journal of Plants Sciences 155, 763-771. Hansen, H.V., 1991. Phylogenetic studies in Compositae tribe Mutisieae. Opera Botanica 109, 150.  79  Hansen, H.V., 1992. Studies in the Calyceraceae with a discussion of its relationship to Compositae. Nordic Journal of Botany 12, 63-75. Irwin, H.S., Barneby, R.C., 1982. The American Cassiinae. Memoirs of the New York Botanical Garden 35, 1-918. Kay, Q.O.N., 1988. More than the eye can see: the unexpected complexity of petal structure. Plants Today, 109-114. Kay, Q.O.N., Daoud, H.S., Stirton, C.H., 1981. Pigment distribution, light reflection and cell structure in petals. Botanical Journal of the Linnean Society 83, 57-84. Kenicer, G.J., Kajita, T., Pennington, R.T., Murata, J., 2005. Systematics and biogeography of Lathyrus (Leguminosae) based on internal transcribed spacer and cpDNA sequence data. American Journal of Botany 92, 1199-1209. Kevan, P.G., Lane, M.A., 1985. Flower petal microtexture is a tactile cue for bees. Proceedings of the National Academy of Sciences USA 82, 4750-4752. Kolossova, N., Shermon, D., Karlson, D., Dudareva, N., 2001. Cellular and subcellular localization of S-adenosyl-L-methionine: benzoic acid carboxyl methyltransferase, the enzyme responsible for biosynthesis of the volatile ester methylbenzoate in snapdragon flowers. Plant Physiology 126, 956-964. Lavin, M., Herendeen, P.S., Wojciechowski, M.F., 2005. Evolutionary rates analysis of Leguminosae implicates a rapid diversification of lineages during tertiary. Systematic Botany 54, 575-594. Lewis, G.B., Schrire, B., Mackinder, B., Lock, M., 2005. Legumes of the world. Kew Publishing, London. Maddison, D.R., Maddison, W.P., 2000. MacClade 4: analysis of phylogeny and character evolution, version 4.0. Sinauer, Sunderland, Massachusetts. 80  Maddison, W.P., Maddison, D.R., 2009. Mesquite: a modular system for evolutionary analysis. Marazzi, B., Endress, P.K., 2008. Patterns and development of floral asymmetry in Senna (Leguminosae, Cassiinae). American Journal of Botany 95, 22-40. Marazzi, B., Endress, P.K., de Queiroz, L.P., Conti, E., 2006. Phylogenetic relationships within Senna (Leguminosae, Cassiinae) based on three chloroplast DNA regions: patterns in the evolution of floral symmetry and extrafloral nectaries. American Journal of Botany 93, 288-303. Mathur, J., 2006. Local interactions shape plant cells. Current Opinion in Plant Biology 18, 4046. McMahon, M., 2005. Phylogenetic relationships and floral evolution in the papilionoid legume clade Amorpheae. Brittonia 57, 397-411. McMahon, M., Hufford, L., 2004. Phylogeny of Amorpheae (Fabaceae: Papilionoideae). American Journal of Botany 91, 1219-1230. McMahon, M., Hufford, L., 2005. Evolution and development in the amorphoid clade (Amorpheae: Papilionoideae: Leguminosae) petal loss and dedifferentiation. International Journal of Plant Sciences 166, 383-396. Noda, K., Glover, B.J., Linstead, P., Martin, C., 1994. Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor. Nature 369, 661664. Pardo, C., Cubas, P., Tahiri, H., 2004. Molecular phylogeny and systematics of Genista (Leguminosae) and related genera based on nucleotide sequences of nrDNA (ITS regions) and cpDNA (trnL-trnF intergenic spacer). Plant Systematics and Evolution 244, 93-119.  81  Payne, T., Clement, J., Arnold, D., Lloyd, A., 1998. Heterogeneous myb genes distinct from GL1 enhance trichome production when overexpressed in Nicotiana tabacum. Development 126, 671-682. Prenner, G., 2004a. Floral development in Polygala myrtifolia (Polygalaceae) and its similarities with Leguminosae. Plant Systematics and Evolution 249, 67-76. Prenner, G., 2004b. Floral ontogeny in Calliandra angustifolia (Leguminosae: Mimosoideae: Ingeae) and its systematic implications. International Journal of Plant Sciences 165, 417426. Prenner, G., Klitgaard, B., 2008. Towards unlocking the deep nodes of Leguminosae: floral development and morphology of the enigmatic Duparquetia orchidacea (Leguminosae, Caesalpinioideae). American Journal of Botany 95, 1349-1365. Ree, R.H., Citerne, H.L., Lavin, M., Cronk, Q.C.B., 2004. Heterogeneous selection on LEGCYC paralogs in relation to flower morphology and the phylogeny of Lupinus (Leguminosae). Molecular Biology and Evolution 21, 321-331. Stirton, C.H., 1981. Petal sculpturing in Papilionoid Legumes. In: Polhil, R.M., Raven, P.H. (Eds.), Advances in legume systematics. HSMO, London, pp. 771-788. Tucker, S., 2002. Floral ontogeny of Cercis (Leguminosae: Caesalpinioideae: Cercideae): does it show convergence with papilionoids? International Journal of Plant Sciences 163, 75-87. van Houwelingen, A., Souer, E., Spelt, K., Kloos, D., Mol, J., Koes, R., 1998. Analysis of flower pigmentation mutants generated by random trasposon mutagenesis in Petunia hybrida. The Plant Journal 13, 39-50. Wang, Z., Luo, Y., Li, X., Wang, L., Xu, S., Yang, J., Weng, L., Sato, S., Tabata, S., Ambrose, M., Rameau, C., Feng, X., Hu, X., Luo, D., 2008. Genetic control of floral zygomorphy in  82  pea (Pisum sativum L.). Proceedings of the National Academy of Sciences of the United States of America 105, 10414-10419. Whitney, H.M., Kolle, M., Andrew, P., Chittka, L., Steiner, U., Glover, B.J., 2009. Floral iridescence, produced by diffractive optics, acts as a cue for animal pollinators. Science 323, 130-133. Wojciechowski, M.F., Lavin, M., Sanderson, M.J., 2004. A phylogeny of Legumes (Leguminosae) based on analyses of the plastid matk gene resolves many well-supported subclades within the family. American Journal of Botany 91, 1846-1862.  83  4 The origin of bird pollination in Macaronesian Lotus (Loteae, Leguminosae)1  4.1 Introduction Macaronesia consists of five volcanic archipelagos (Azores, Madeira, the Salvage Islands, the Canary Islands and Cape Verde) as well as a region of African mainland (southern Morocco and the former Spanish West Africa) known as the ―Macaronesian enclave‖ (Fig. 4.1). The Macaronesian flora has a high degree of endemism, 20% overall (Humphries, 1979) and 40% in the Canary Islands alone (Santos-Guerra, 1999). It also has a diverse range of altitudinal zones (0-3200m), island ages (ranging from 0.8 to 21 Ma) and distances from main continental areas of Europe or Africa, ranging from 95 to 1600 km (Carracedo, 1994; Carracedo et al., 2002; Humphries, 1979). Bird pollination has evolved in this region in the lineages of at least eleven endemic plant species from six genera, Canarina and Musschia (Campanulaceae), Isoplexis (Scrophulariaceae), Echium (Boraginaceae), Lotus (Fabaceae) and Lavatera (Malvaceae). These species possess several features associated with opportunistic nectar-feeding birds, including red-orange flowers, abundant dilute nectar, diurnal anthesis, extended flower life span, and loss of scent and landing platform (Dupont et al., 2004; Olesen, 1985, 1988; Olesen and Valido, 2003a; Ollerton et al., 2009; Rodriguez and Valido, 2008; Valido et al., 2004).  1  A version of this chapter will be submitted for publication: Ojeda, I. Santos-Guerra, A., Oliva-  Tejera, F. Jaen Molina, R. Caujape, J. Marrero, A. and Cronk. Q. C. The origin of bird pollination in Macaronesian Lotus (Loteae, Leguminosae)  84  Recent studies have suggested at least five passerine birds, the Canarian Chiffchaff (Phylloscopus canariensis), the Blue tit (Parus caeruleus), the Sardinian warbler (Sylvia melanocephala), canary (Serineus canarius) and chaffinch (Fringilla coelebs), as the main pollinators of these species (Ollerton et al., 2009; Rodriguez and Valido, 2008; Sletzer, 2005; Valido et al., 2002). Despite the relatively large number of lineages with a bird pollination syndrome in this region, to date there are only two studies that demonstrate the role of these opportunistic nectar feeders as effective pollinators: in Isoplexis and Canarina (Ollerton et al., 2009; Rodriguez and Valido, 2008). In the remaining groups only foraging observations have been reported. Several hypotheses have been suggested to explain the origin and maintenance of this pollination syndrome in Macaronesia: 1) the ―de novo specialist‖ hypothesis states that presumably extinct specialist nectarivorous birds on the islands exerted selective pressures on the flowers, followed by their further maintenance by opportunistic birds after the extinction of the specialist birds (Olesen, 1985; Vogel et al., 1984); 2) the ―relict‖ hypothesis suggests that the selection and evolution took place in mainland areas before the plant taxa colonized the islands, once in the islands the specialist bird was replaced by non-specialist nectarivorous passerines (Valido et al., 2004); 3) the ―de novo opportunistic‖ hypothesis suggests that current opportunistic birds acted as selective agents on the floral traits on the islands (Dupont et al., 2004; Valido et al., 2004). To date, evidence from Lavatera and Canarina (Fuertes-Aguilar et al., 2002), and Isoplexis (Bräuchler et al., 2004; Rodriguez and Valido, 2008) suggests that these floral features are plesiomorphic on the island and may be relictual. It is therefore likely that the evolution of these flower features occurred on the mainland before the colonization of Macaronesia. However, in the remaining cases (Lotus, Musschia and Echium) this is still unresolved (Valido et 85  al., 2004) and it is unknown when these flower features evolved within these groups. Birdpollinated Echium (Bohle et al., 1996; Dupont and Skov, 2004; Valido et al., 2002) and Lotus (Allan et al., 2004) have derived positions within entomophilous clades and this may indicate a recent origin in the Canary Islands. However, in Lotus it has not yet been possible to determine unequivocally the phylogenetic origin of bird pollination or the most closely related entomophilous species, due to incomplete sampling and the low resolution of the internal nuclear ribosomal transcribed spacer region (ITS) used in previous studies (Allan et al., 2004; Degtjareva et al., 2006). Bird pollination occurs in four species within Macaronesian Lotus (Olesen, 1985; Ollerton et al., 2009; Valido et al., 2004). These four species are placed in a group commonly referred to as section Rhyncholotus (Monod) D.D. Sokoloff (Degtjareva et al., 2006; Olesen, 1985; Sandral et al., 2006; Valido et al., 2004). They have several floral traits associated with the bird pollination syndrome. These traits include: large flowers, large quantities of dilute nectar (mainly composed of hexose sugars), red-orange flower color, long lived flowers, upward orientation of the flower, and change in the relative size and shape of petals (Dupont et al., 2004; Olesen, 1985; Ollerton et al., 2009; Valido et al., 2004). The syndrome is associated with pollination by opportunistic nectar-feeders. Two birds, the Canarian chiffchaff and the blue tit, have been reported foraging in cultivated individuals of at least two of these plant species (Ollerton et al., 2009; Sletzer, 2005). These four bird-pollinated species seem to have a derived position within the entomophilous species-rich section Pedrosia (Lowe) Christ (Allan et al., 2004) . The species of sections Pedrosia and Rhyncholotus represent an example of an island radiation within Macaronesia. The group comprises about 40 species of mainly perennial herbs with a high diversity of vegetative and flower features. The main diagnostic feature common to 86  both sections (Pedrosia s.l.) is the presence of a tooth (also mentioned as forked style) on the ventral side of the style (Kramina and Sokoloff, 1999). This flower feature has been used to distinguish Pedrosia s.l. from other sections within Lotus. Members of Pedrosia s.l. occur in a wide range of habitats ranging from rocky coast to mountain pine forest at an elevation above 1200 m. They are represented in all five archipelagos as well as in mainland Africa and Europe (Sandral et al., 2006) (Fig. 4.1 and 4.2). Pedrosia s.l. is also characterized by a high degree of endemism; some of its species are restricted to specific habitats within one island. At least seven species within the group are considered critically endangered (CE), one species endangered (E) and one vulnerable (V) (Bañares et al., 2004). Taxonomically, both groups have been variously united as Pedrosia s. l. (Rhyncholotus + Pedrosia), or considered as separate sections (Brand, 1898), subgenera (Bentham, 1865; Monod, 1980) or genera (Christ, 1888; Kunkel, 1974). The only characteristics separating Rhyncholotus from Pedrosia are those associated with floral adaptations to bird pollination (Perez de Paz, 1990; Sandral et al., 2006). Currently, Pedrosia and Rhyncholotus are recognized as a two separate sections within Lotus (Degtjareva et al., 2006), the former with about 36 species and the latter with four species. Previous analyses have not completely clarified their relationships and the current phylogenetic evidence suggests that Pedrosia is paraphyletic, with Rhyncholotus embedded in one of the clades within Pedrosia (Allan et al., 2004; Degtjareva et al., 2008; Degtjareva et al., 2006). The addition of morphology to the nrITS phylogeny in the most recent analysis of the group recovered Rhyncholotus as monophyletic (Degtjareva et al., 2006) within a paraphyletic Pedrosia, but the relationships within Pedrosia s.l. are generally still poorly resolved. Given that the four Rhyncholotus species are embedded within Pedrosia, I will refer to the four birdpollinated species as the ―rhyncholotus group‖. 87  This lack of resolution may be partly due to an incomplete sampling of the group, and partly because this group is a recent island radiation with low levels of variation at the nrITS locus. This lack of resolution in the current phylogenetic analyses has hindered taxonomic decisions (Sandral et al., 2006) and more accurate interpretations of island colonization (Allan et al., 2004). This limitation in resolution has also prevented further studies of the evolution of bird pollination and traits associated with this pollination syndrome transition (Allan et al., 2004; Dupont et al., 2004; Sandral et al., 2006).  4.1.1 Objectives of the study My aims in this chapter are (i) to clarify the phylogenetic relationships within Pedrosia s.l. with special regard to determining the sister-group relationship of the bird-pollinated species, (ii) to test the monophyly of the bird-pollinated group, and (iii) to assess the timing and ecological context of the origin of bird pollination in this group.  4.2 Materials and methods 4.2.1 Taxon sampling Here, I analyze a nested series of three data sets with different sample numbers and gene regions. (1) ITS analysis. First I made a preliminary analysis with a very comprehensive sampling using the nuclear ribosomal ITS region alone. This represented, with at least one sample, almost every taxon or population that has been recognized at the species level with this group. I also included samples from populations that may represent new species based on morphological and/or genetic evidence (Oliva-Tejera et al., 2005; Oliva-Tejera et al., 2006). This preliminary sampling included 125 samples in total, 118 from the ingroup, representing 37 described species, 88  two subspecies, one variety and four undescribed new species. Of these samples, 92 are new sequences generated in this analysis. In this analysis I also included 26 sequences from previous phylogenetic analyses deposited in GenBank (Allan et al., 2004; Allan et al., 2003; Degtjareva et al., 2008; Degtjareva et al., 2006) (Table 4.2). I was able to include at least two samples each for 31 of the 41 species considered (75%) with the purpose of having a good representation of the geographical distribution of some species and also being able to explore the potential intraspecific sequence variation of this gene region, a feature that has been suggested previously in L. creticus (Allan et al., 2004; Degtjareva et al., 2006; Sandral et al., 2006). This analysis allowed all species that are broadly related to section Rhyncholotus to be detected and used in subsequent analyses. (2) Four-gene analysis. As a result of the previous analysis a subset of samples was selected for more detailed analysis with more gene regions. I included 54 samples (all newly collected for this analysis) representing 39 species, two subspecies, three varieties and one undescribed species within Pedrosia s.l. (Table 4.1). This data set was analyzed with four gene regions, one nuclear (ITS) and three plastid gene regions (matK, trnH-psbA and CYB6). This data set was analyzed both separately (nuclear or plastid only) and in combination. (3) Six-gene analysis. Finally, I analyzed a data set of 21 samples (19 from the ingroup) with six gene regions, four nuclear (ITS, LjCYC1, 2 and 3) and two plastid (trnH-psbA and matK). CYB6 was not use in this analysis due to its lack of variability. LjCYC1, 2 and -3 are developmental genes (transcription factors of the TCP family) for which further details are given in Chapter 6. This data set was used to determine the closest relatives of the bird-pollinated species. It comprises 15 species and included several accessions of the closest relatives that I identified in previous analyses (Table 4.1).  89  Only three species, L. chazalei, L. tibesticus and L. loweanus were not available for the analyses described above and were not included. Lotus chazalei is distributed in the coastal region of Mauritania, extreme SW Sahara and extreme SW Morocco, L. tibesticus is an endemic from Chad and L. loweanus is an endemic to the Island of Porto Santo, Madeira (Fig. 4.1). I do not believe that these are critical species to determine the closest relative of the bird-pollinated species (rhyncholotus group). There is strong morphological evidence that they are closely related to other species in the analysis, L. tibesticus to L. jolyi, L. loweanus to L. argyrodes group and L. chazalei to L. assakensis group, but not to rhyncholotus (Sandral et al., 2006).  Outgroup selection I selected four species from Lotus section Lotus as an outgroup. This group contains the model legume Lotus japonicus and three related species. This group was chosen as some of the gene regions used in this analysis were first isolated and studied in detail in L. japonicus. Additionally, I also included two species from a different genus in the Loteae, Hosackia, which occur only in North America and are much less closely related to the Macaronesian Lotus.  4.2.2 DNA Extraction and sequence data analysis Genomic DNA was extracted from either fresh leaves, silica gel dried leaf material or voucher specimens (Table 4.1 and 4.2) following a modification of the cetyl-trimethylammonium bromide (CTAB) procedure (Doyle and Doyle, 1987). In total I sequenced and analyzed four nuclear and three plastid regions.  90  4.2.3 Nuclear regions The ITS region was amplified (30 cycles of 96 C for 1 min, 49C for 1 min, 72C for 1 min) using primer 4 and 5 (White et al., 1990). I included three homologues of the CYCLOIDEA gene from Lotus japonicus. Lotus japonicus CYCLOIDEA1 (LjCYC1) was amplified (95 C for 2 min, 30 cycles of 94 C for 30 s, 55 C for 1 min and 72 C 2 min, and 72 C for 2 min) using a combination of two specific and two general primers. The region that encompasses the TCP domain was amplified using CYC.1.1F and LEGCYCR1 and the region with the R domain using LEGCYCF and CYC1.1R. LjCYC2 was amplified, either using a combination of two specific primers LC2.1F and LC2.1R (95 C for 2 min, 30 cycles of 94 C for 45 s, 50-57 C for 1 min and 72 C 2 min, and 72 C for 2 min) or a combination of LC2.1F and LEGCYCR for the region that includes the TCP domain and LEGCYCF and LC2.1R for the region that includes the R domain (95 C for 2 min, 30 cycles of 94 C for 45 s, 55 C for 1 min and 72 C 2 min, and 72 C for 2 min). LjCYC3 was amplified with two primer combinations, the region including the TCP domain was amplified (95 C for 2 min, 30 cycles of 94 C for 30 s, 45 C for 1 min and 72 C 2 min, and 72 C for 7 min) using the primers CYC3.2F and LEGCYCR1 (Citerne et al., 2003). The region that contains the R domain was amplified with LEGCYCF and CYC3.1R (Table 4.3).  4.2.4 Plastid regions The cytochrome B6 (CYB6) was amplified (94 C for 2 min, 35 cycles of 94 C for 20 s, 55 C for 20 s and 72 C 2 min, and 72 C for 5 min) using the primer CYB6-F and CYB6-R (Choi et al., 2006b). The intergenic region trnH-psbA (Kress et al., 2005) and the matK gene  91  were amplified (94 C for 3 min, 35 cycles of 94 C for 30 s, 45 C for 1 min and 72 C 2 min, and 72 C for 2 min) using the primers matKX and matK3.2, respectively (Table 4.3). Each locus was amplified, sequenced and the raw sequence data were imported to Sequencher 4.1 for editing and combining of contig sequences. Consensus sequences were imported to Se-Al ver. 1.0 (Rambaut, 1996) and aligned manually using conserved regions in order to identify homologues sequences among species. Gaps in the ITS, trnH-psbA, and LjCYC1, 2 and 3 were codified as characters (Simmons and Ochoterena, 2000).  4.2.5 Phylogenetic analyses Phylogenetic analyses were performed using maximum parsimony (MP), maximum likelihood (ML) and the Bayesian method. Parsimony analysis was conducted using PAUP* ver. 4.0b10* (Swofford, 2001) assuming unordered characters and equal character weighting. Heuristic searches were performed with 10000 random stepwise-addition replicates TBR branchswapping, and MULTREES optimization was used. Consistency index (excluding uninformative characters) and retention index were also calculated (Farris, 1989; Kluge and Farris, 1969). Branch support was analyzed with bootstrap (Felsenstein, 1985) using a simple addition sequence and 10 000 replicates with all parameters similar as used in the MP strategy. Additionally, I also analyzed the same data sets using maximum likelihood as implemented in Garli Version 1.0 (Zwickl, 2006) and Bayesian analysis using Mr. Bayes (Huelsenbeck and Ronquist, 2001; Ronquist and Huelsenbeck, 2003). The best-fit model for each region was determined using the AIC as implemented in jMODELTEST ver. 0.1.1. (Posada, 2008). I used the GTR +G + I; lset nst = 6 rates = invgamma and the HYK models. Bayesian analyses were run using four Monte Carlo chains, a random tree as a starting point, sampling every 1, 000 generations, and continuing for 5,000,000 generations. ML analyses with 92  Garli were run using the same two models mentioned above with stepwise staring tree topology, two search rep and 100 bootstrap.  4.2.6 Dating the origin of bird pollination Divergence times within the Macaronesian Lotus were estimated using the program Beast v1.5.4 (Drummond and Rambaut, 2007), which estimates branch lengths, topologies, substitution parameter models and dates simultaneously. I ran this analysis using two data sets, a combined matrix of 52 samples and four gene regions (ITS, matK, trnH-psbA and CYB6) with a total of 2092 bp, and a data set of six genes (ITS, LjCYC1, 2 and 3, matK, trnH-psbA and CYB6) and 21 samples. I used a constant-rate Yule (speciation process) prior and all other priors and operators were the default settings. Four independent runs were performed using the uncorrelated lognormal relaxed-clock model (Drummond et al., 2006) for 50,000,000 generations. Trees and parameters were sampled every 5,000 generations yielding a total of 10,000 trees, with a burn in of 5, 000,000. All analyses were run using the HYK+gamma and GTR +G + I; lset nst = 6 rates = invgamma substitution model. The Beast file was created using the BEAUti program v 1.5.4 (part of the program BEAST). The performance of each run was further analyzed with the program Tracer. Mean parameter estimates and 95% highest posterior probabilities (HPDs) were determined by analyzing the Beast tree files with TreeAnnotator v 1.5.4 (Drummond and Rambaut, 2007). Trees were visualized and edited with Figtree v1.3.1. This analysis was constrained with the best hypothesis of relationship (topology) of this group obtained from MP and ML in previous analyses. Macaronesia lacks reliable fossil records and instead I used well-known geological estimates of maximum island age as my calibration points (Fig. 4.1 and 4.2). For this particular analysis I used three calibration points. This group has several endemic species on different 93  islands and I used the distribution of two endemics, L. sessilifolius subsp. villossisimus (El Hierro, 1.12 Ma) and L. sessilifolius subsp. sessilifolius (La Palma 1.77 Ma) as calibration points (Ancochea et al., 1994; Carracedo, 1994). Both species are endemic to these islands and they were not able to colonize and diversify previous to the emergence of these two islands. The ages of La Palma and El Hierro have been used previously as calibration points in other plants groups (Kim et al., 2008b; Percy et al., 2004). The third calibration point of 21 Ma was based on the age of the oldest island, Fuerteventura, as an upper limit for the colonization of the Canary Islands (Carracedo, 1994) and therefore an upper limit for the age of the most common recent ancestor (MRCA) for the species of this archipelago. This approach of using multiple maximum ages has been used in several analyses within this archipelago, both in animals and plants (Anderson et al., 2009; Cox et al., 2010; García-Maroto et al., 2009; Kim et al., 2008b; O'Leary, 2009), and it has been suggested that is less prone to inaccurate date estimations (Anderson et al., 2009).  4.3 Results 4.3.1 Major clades of Pedrosia s.l. in Macaronesia My analysis using the large data matrix of 125 ITS accessions recovered four major clades within the Macaronesian Lotus section Pedrosia and the rhyncholotus group (Fig. 4.3, Clades A-D), and no differences were observed when GenBank sequences where excluded or included from this analysis, The same topology was recovered with the data set of 54 accessions. Clade A contains all the Cape Verde species, which form a monophyletic group together with one African lineage of three species. Clade B includes species from the Canary Islands, Madeira and the Azores. The Madeiran and the Azorean species are located in an early divergent position within this clade together with samples of an undescribed species from Tenerife. The four birdpollinated species (rhyncholotus) are all in clade B, and are unresolved with other five species of 94  Pedrosia s.s. referred to as the ―Lotus sessilifolius group‖. The level of variation within the ITS phylogeny does not indicate unequivocally whether the bird-pollinated species are monophyletic or shed any light on the exact sister group relationship of rhyncholotus (Fig. 4.3). This lack of resolution is despite the fact that the ITS nuclear region is more variable that the plastid regions (Table 4.4). Clade C also includes species mainly from the Canary Islands, one species from the Salvage Islands, one from Madeira and three African species. Two of these African species form a sister group of the remaining species within this clade and the remaining species are divided into two other subclades. Of these, subclade I includes exclusively Canary Island species with a mountain distribution, whereas subclade II included one species from Madeira and another species restricted to the Salvage Islands. Finally, another early divergent clade consists of an African species (Lotus jolyi) (clade D). The inclusion of indels (two in the ITS matrix for section Pedrosia s.l.) in the analysis increased the resolution within clade C. No additional improvement in resolution was observed in other clades. Less resolved topologies were recovered with the separate analyses of the three plastid regions matK, trnH-psbA and CYB6, and even the combined analysis of the three plastid regions resulted in a less resolved and well-supported phylogeny than the ITS analysis alone (Fig. 4.4). This combined plastid analysis recovered only one group (clade A) observed with the ITS phylogeny and there was an incongruence of the placement of four species (L. callis-viridis, L. kunkelii, L. arinagensis and L. emeroides) between clades B and C. This incongruence is probably due to ―soft‖ incongruence, given the low levels of sequence variation and the lack of resolution within the three plastid regions. Only one ―hard‖ incongruence was detected in the Cape Verdean clade (clade A), in which contrasting topologies in the two data sets have moderate bootstrap support. However, importantly for this study, the position of the four bird95  pollinated species was consistent in all analyses and the same species, L. sessilifolius and L. mascaensis, were recovered in both data sets as the closest relatives of the ornithophilous species. The combined analysis of these four regions (ITS, trnH-psbA, matK and CYB6) resulted in a more highly resolved phylogeny with the recovery of the same major clades as in the ITS analysis. The position of the four bird-pollinated species is the same, whether in separate analyses (nuclear and plastid), or in combination. The four bird-pollinated species are placed together within clade B, and they are revealed to be closely related to two species of the L. sessilifolius group, L. sessilifolius and L. mascaensis, and one undescribed species from Tenerife (Fig. 4.5). However, there is still a lack of resolution of the relationships among these species and it is still not possible to unequivocally identify the sister group of the bird-pollinated species. The above results are all based on maximum parsimony analyses. However, similar results were obtained when the same data sets were analyzed with maximum likelihood or Bayesian analyses, and only minor discrepancies were observed within the four major clades obtained from MP analyses. The only major difference was the placement of L. jolyi. All MP analyses placed this species as the sister group of all Macaronesian Lotus, while Bayesian analysis place it as a sister clade of the Cape Verde clade (clade A) and ML placed it as the sister species of clades B and C. However, in all analyses the support for the placement of this species is low and further studies should be considered to address its position.  4.3.2 Recovery of the sister group of rhyncholotus using a six gene region analysis with the inclusion of CYCLOIDEA homologues The previous combined analysis of four regions identified two species (L. sessilifolius and L. mascaensis) as the closest relatives of the bird-pollinated species within clade B. A further 96  analysis of a targeted data set using three additional nuclear regions and two plastid regions was therefore undertaken. One plastid region used in previous analyses (CYB6) was abandoned as it had no variation at this level. The addition of three CYCLOIDEA homologues increased the resolution within this clade. The three CYCLOIDEA homologues gave similar topologies when analyzed separately and in combination using either MP, ML or Bayesian analyses. Among these three paralogues, LjCYC1 seems to be the most variable copy (Table 4.5). For this sample set the combined analysis of ITS and the two plastid regions (matK and trnH-psbA) gave less resolution than the three CYCLOIDEA copies together, and most of the resolution within clade B in this analysis was obtained from the three CYCLOIDEA homologues. The combined analysis of the six regions increased the resolution and indicates that L. sessilifolius is likely the closest relative of the bird-pollinated species. I sampled representatives from all of the geographic distribution (samples from four islands) and intraspecific taxonomic groups (two subspecies and one variety) within this species and it appears that L. sessilifolius s.l. is likely a paraphyletic species with respect to the rhyncholotus group. However, with these data sets I was unable to identify unequivocally the population or populations closest to the four birdpollinated species (Fig. 4.6A-B).  4.3.3 Dating the origin of bird-pollinated Lotus in the Canary Islands Calibration of the Macaronesian Lotus indicates that this group likely colonized this region in three events between 6 to 4.3 Ma between late Miocene and early Pliocene. Under this scenario the earliest colonization occurred in the Cape Verdean archipelago, circa 6.1 (0.50-7.29) Ma (Clade A) with the latest in the Canary Island, circa 4.3 Ma (0.58-6.38) (Fig. 4.9). The four species with the bird pollination syndrome (rhyncholotus group) appears to be of recent origin with the most recent common ancestor (MRCA) for L. sessilifolius and the 97  rhyncholotus group dated at 2.2 Ma and an age of for the bird-pollinated rhyncholotus species (MRCA) of 1.2 Ma (Fig. 4.7). The six gene chronogram (Fig. 4.10) gives an age (MRCA) for the L. eremiticus/L. pyranthus group of 0.67 Ma and of L. berthelotii/L. maculatus of 1.11 Ma, and a date of the whole L. sessilifolius group, including rhyncholotus, of 2.09 Ma (Table 4.6). The evolution of bird pollination is therefore associated with a recent radiation of Macaronesian Lotus (the L. sessilifolius group) that started the last 2 Ma in the islands of Tenerife, La Palma, El Hierro and to a lesser extent in Gran Canaria around the beginning of the Pleistocene.  4.4 Discussion 4.4.1. Closest relative of the four bird-pollinated species A previous study has suggested either species of the L. argyrodes group from Madeira and the Azores (4 spp) (Fig. 4.7A-C) or members of the L. sessilifolius group (5 spp) (Fig. 4.7DI) as the most closely relatives of the bird-pollinated rhyncholotus (Sandral et al., 2006). Most members of these two groups have relatively larger flowers compared to the rest of the Pedrosia s.s. and both groups have the ability to modify flower colour after anthesis by producing anthocyanidins similar to those observed in red-orange, bird-pollinated flowers (see Chapter 7). However, previous phylogenetic analyses were based on less comprehensive sampling in both of these two groups (Allan et al., 2004; Degtjareva et al., 2006) and the determination of the closest relative of rhyncholotus remained unclear. My data set of four gene regions (ITS, trnH-psbA, matK and CYB6) (Fig.4.4) indicates that three species, L. mascaensis, L. sessilifolius and Lotus sp. nov. 1 from Tenerife are likely the closest relatives of the rhyncholotus group. The first two species, L. mascaensis, L. sessilifolius, share some morphological features with the bird-pollinated species. Both species have flowers in 98  a somewhat upward orientation, an intermediate inclination between the rest of Pedrosia s.s. and the rhyncholotus group. Additionally, the leaflets of both species are filiform or linear (Fig. 4.7D-I), a feature present also in rhyncholotus (Fig. 4.8). In contrast, Lotus sp. nov. 1 does not share these features, the leaflets are oblong to lanceolate and the flowers are held in a horizontal position, a feature commonly observed in the rest of Pedrosia s.s. This undescribed species shares leaflet morphology with the L. argyrodes group (4 spp.) (Fig. 4.7A-C) and L. emeroides from La Gomera, in all of which the three distal leaflets are longer than the two basal. My data set of six gene regions (ITS, LjCYC1, 2 and 3, trnH-psbA, matK and CYB6) specifically supports L. sessilifolius as the closest relative of the four bird-pollinated rhyncholotus (Fig. 4.6A and B). This species exits as a complex group which is distributed in four islands (La Palma, La Gomera, Tenerife and El Hierro) within the Canary Island archipelago (Sandral et al., 2006) and it is the most widely distributed species in Lotus subg. Pedrosia (Fig. 4.2). Lotus sessilifolius has been divided into three subtaxa (two subspecies and one variety), L. sessilifolius subp. sessilifolius from Tenerife, La Gomera and La Palma, L. sessilifolius subsp. villosissimus from El Hierro (Sandral et al., 2006) and L. sessilifolius var. pentaphyllus from Tenerife (Fig. 4.7G-I) (Bramwell and Bramwell, 2001). According to the six gene data set, L. sessilifolius represents a paraphyletic group from which rhyncholotus evolved. Similar topologies were obtained with ML and Bayesian analyses and the position of this species as the closest relative of the rhyncholotus group is well supported in these analyses (Fig. 4.6AB). The relationship between individual accessions of this species and rhyncholotus is not unequivocally resolved, but in all analyses I recovered L. sessilifolius var. pentaphyllus as the earliest diverging accession within this group (Fig. 4.6A and B). My results from the six gene data set raises the possibility that bird pollination may have a double origin (Fig. 4.6A and B), although there is no statistical support for this, and on 99  morphological grounds a single origin would seem to be more likely because of the many striking similarities between all four bird-pollinated species (Fig. 4.8). Further analyses are required to fully determine whether bird pollination has a single or double origin in this particular group.  4.4.2. How old is the bird pollination syndrome in Macaronesian Lotus? The phylogenetic results reported here, support the scenario that the floral features associated with bird pollination in Macaronesian Lotus (rhyncholotus group) evolved relatively recent. All my phylogenetic analyses place the four species of this group in a highly derived position within the section Pedrosia, apparently nested within a single species (Fig. 4.6A-B). All my dating analyses strongly suggest that the clade containing the four rhyncholotus species evolved within the last 2 Ma representing a recent island radiation, postdating the Tertiary-Quaternary boundary (2.6 Ma). Similar age estimates were obtained when rhyncholotus was constrained to be monophyletic in both data sets (Table 4.4). The four rhyncholotus species seem to have shared a MRCA with L. sessilifolius the last 2.08-2.70 Ma. The four rhyncholotus species are therefore a neoendemic lineage (Cronk, 1992; Vargas, 2007) of recent evolution in this archipelago. It is truly remarkable that a clade which has been considered in the past a separate genus should in actuality have evolved so recently (and in some analyses be apparently nested within a single extant species, L. sessilifolius). These results, however, should be considered with caution. I calibrated the phylogenetic trees using the ages of the islands as upper bound limits and the current distribution of endemics species. It is possible, although unlikely, that these species could have been more widespread distributed in the past and their current distribution does not reflect previous historic ranges. Therefore, using the current distribution on endemics might underestimate the ages obtained. 100  Bird pollination in Echium likely occurred within the same time window as in Lotus. The bird-pollinated E. wildpretii seems to have diverged in the last 1 Ma from the melittophilous E. pinianana (García-Maroto et al., 2009). In contrast, bird pollination in Lavatera, Canarina and Isoplexis probably evolved earlier, as they represent apparently older lineages (Bräuchler et al., 2004; Fuertes-Aguilar et al., 2002; Rodriguez and Valido, 2008); although no phylogenetic dating studies have been carried out in these groups. The age estimates reported here for Lotus together with the Echium estimate are the only age estimations for the evolution of bird pollination in Macaronesia. Outside Macaronesia, much older dates have been inferred for the evolution of bird pollination in various groups studied, such as Alloxylon, Telopea and Embothrium (Proteaceae), where bird pollination apparently originated in the early Eocene (52.8 Ma ± 6.8) (Barker et al., 2007). Under this scenario of recent evolution, it is likely that these floral features evolved in the Canary Islands Lotus under the selective pressure of opportunistic passerine birds (―de novo opportunistic‖ hypothesis) (Dupont et al., 2004; Valido et al., 2004). There are records of floral visits by two passerine birds, the Canarian chiffchaff and the blue tit (Ollerton et al., 2009; Sletzer, 2005) in this group. However, there has been no rigorous analysis of the effectiveness of these two bird species as pollinators in the four species of rhyncholotus. In fact, these observations have only been made in cultivated populations, and only in one species, L. berthelotii and its hybrid with L. maculatus. There are no records of bird visits to the other three species of rhyncholotus, either in cultivation or in individuals from wild populations. Therefore, the role of these birds as selective agents for the evolution of these floral traits remains largely uninvestigated, and this is complicated by the rarity of the plants in the wild. Age colonizations of some passerines in Macaronesia (including Parus, Serineus and Sylvia) have been estimated 101  between 0.008 to 2.3 Ma (using divergence times calculated after rate estimates of 2 % sequence divergence per Ma) (reviewed in Dietzen, 2007). In particular, the polytipic Parus caeruleus contains at least 15 subspecies distributed from Macaronesia, Europe and Africa. Four endemic subspecies are distributed in the Canary Island archipelago, P. caeruleus subsp. ombriosus (El Hierro), P. caeruleus subsp. palmensis (La Palma), P. caeruleus subsp. teneriffae (La Gomera, Tenerife, Gran Canaria) and P. caeruleus subsp. degener (Fuerteventura, Lanzarote), that colonized and radiated within this archipelago within the last 0.3- 1.5 Ma (Dietzen, 2007). This passerine group seems to have radiated more or less at the same time as the rhyncholotus group. At least one of the putatively pollinators of rhyncholotus seems to have recently colonized and radiated in this archipelago, thus supporting the de novo opportunistic explanation for the origin of this pollination syndrome. Another explanation is that specialist bird were once distributed in the Canary Island archipelago, and after their extinction, opportunistic passerine birds might have occupied their niche. However, there is no current fossil evidence to support this. All four species of rhyncholotus are considered endangered in the wild with much reduced populations (Caceres et al., 2004a, b; Gómez and Coello, 2004; Ojeda and Marrero, 2004), which is partially responsible for the lack of more comprehensive analyses. What is then the selective advantage of bird pollination in rhyncholotus, and the other plant lineages in Macaronesia? Birds are more reliable pollinators under harsh environmental conditions and /or unpredictable environment, such as high altitudes (Stiles, 1978). Birds also have a higher range of mobility than insects, and therefore are more effective pollinators over long distances in isolated populations. It has been shown in Penstemon (Plantaginaceae) (Kramer et al., 2011) and Streptocarpus (Gesneriaceae) (Hughes et al., 2007a) that bird-pollinated species possess a greater functional connectivity among distant populations than bee-pollinated species. Bird-pollinated species have less genetic structure that their bee-pollinated counterparts, 102  evidence of more effective gene flow among isolated distant populations. The four birdpollinated rhyncholotus have small, isolated populations, and the foraging behaviour of birds would increase population connectivity.  4.4.3. The availability of new niches and the evolution of ornithophily in Lotus The current distribution of the four bird-pollinated species suggests that this syndrome may have evolved as new habitats became available due to recent volcanic activity and the lowering of the relative sea level in the islands of Tenerife, El Hierro and La Palma. Lotus berthelotii and L. maculatus are distributed in the central parts of Tenerife (Gómez and Coello, 2004; Hind, 2008; Martín and Berriel, 2007; Ojeda and Marrero, 2004), which were formed in the last 3 Ma, when volcanic activity united the older regions of Anaga, Teno and Adeje (Carracedo et al., 2002). No recorded observations exist of populations of these two species in these three older regions in Tenerife. The other two species, L. pyranthus and L. eremiticus are endemics of La Palma (Caceres et al., 2004a; Coello, 2007; Gonzáles et al., 2004; Medina, 2008), an island that emerged in the last 1.7 Ma (Carracedo, 1994). However, it must be borne in mind that their current distributions may not reflect their original distribution, either in the geological past, or before European settlement and the increase of agricultural activity (and, more recently, development for tourism). All four species were described relatively recently. Lotus berthelotii was the first species described of the group in 1881 by Masferre-Arquimbau (Masferrer-Arquimbau, 1881) and the last species, L. pyranthus was described two decades ago (Perez de Paz, 1990). Therefore, we have virtually no information on the distribution of these plants prior to the twentieth century. Compared to radiations driven by key morphological innovations, such as spur length in columbines (Hodges, 1997), the availability of novel habitats may have been a more important 103  diversification factor for this group, and consequently for the evolution of bird pollination in rhyncholotus. The emergence of new volcanic terrain and new islands may have had effects on bird behaviour that facilitated the evolution of bird pollination. Birds on islands tend to have a reduced interspecific competition compared to those in continental habitats (Crowell, 1962; MacArthur et al., 1972). As a result birds on islands tend to increase their dietary range, and this may have led to the inclusion of nectar as an additional food source (Valido et al., 2004). This phenomenon of niche widening due to relaxed interspecific competition in islands has been also reported in lizards (Olesen and Valido, 2003b). Interestingly, a lizard (Gallotia galloti Oudart, Lacertidae) has also been reported foraging in flowers of L. maculatus and L. berthelotii (Ollerton et al., 2009); however, their role as an effective pollinator has not been fully assessed.  104  Table 4.1 Samples included in the phylogenetic analysis of the two data sets with 21 samples only (*) and with the 54 samples within Pedrosia s.l. JAO= Jardín de Aclimatación de la Orotava, JBCVC= Jardín Botánico Canario “Viera y Clavijo”, UBC= University of British Columbia. T= Tenerife, GC= Gran Canaria, G= La Gomera, P=La Palma, H=El Hierro, CV=Cape Verde, M=Madeira  Taxon Outgroup Hosackia chihuahuana S. Watson Hosackia gracilis Benth. Lotus japonicus MG20 (Regel) K. Larsen *Lotus japonicus Gifu B129 (Regel) K. Larsen Lotus filicaulis Durieu Lotus corniculatus L. Lotus burttii Borsos Ingroup Lotus section Pedrosia (Lowe) Christ. Lotus arborescens Lowe ex Cout. *Lotus arenarius Brot. *Lotus arinagensis Bramwell. *Lotus argyrodes R.P Murray  Lotus assakensis Brand *Lotus azoricus P. W. Ball Lotus bollei Christ [=L. purpureus sensu Sandral et al] Lotus brunneri Webb in Hooker  Lotus callis-viridis Bramwell & D.H. Davis Lotus campylocladus Webb & Berthel. Lotus creticus L. Lotus dumetorum Webb ex R. P. Murray Lotus emeroides R. P. Murray Lotus eriosolen (Maire) Mader & Podlech L. erythrorhyzus Bolle L. genistoides Webb. (nom. nudum)  Collection information  Voucher, herbarium or GenBank  Cultivated UBC # PI 262405 Mexico Cultivated UBC Cultivated UBC  Ojeda 79/UBC Ojeda 69/UBC  Cultivated UBC  Ojeda 70/UBC  Cultivated UBC Vancouver, BC Cultivated UBC seeds from Univ. Miyazaki  Ojeda 71/UBC Ojeda 46/UBC Ojeda 72/UBC  Cultivated JBCVC # 164/06, Morro Cove Roche Sao Nicolao CV Cultivated UBC # PI 631956, Kourigba, Morocco DNA bank # 651 cultivated JBCVC # 138/01 Cultivated JBCVC # 5485B/UPM/07 (Banco Germoplasma UPM, Dr. Gómez Campo) Ponta Pargo, M Voucher, Tarfaya-Tan Tan Sahara, Africa Cultivated JAO # 161-00 Cultivated JBCVC # 163/06, Baia das Gatas on Monte Verde, Sao Vicente, CV Cultivated JBCVC # 514B/07, Mocete Negro (between Stª María and Pedro Lume) Sal, CV DNA Bank # 654, cultivated JBCVC, # 369/04 Arona-Ifonche, T  F. Oliva / A. Marrero / JBCVC / Ojeda 180/ UBC Ojeda 78/UBC  Cultivated JBCVC # 64/05 (# 339/97) Cabo Pino. Mirador Jardina, Mercedes, Anaga, T Teno Alto, Teno, T Inchora, G Cultivated UBC # PI 631959, Ouarzate, Morocco Fuerteventura DNA Bank # 655, cultivated JBCVC # 330/02  B. Navarro /JBCVC / Ojeda 188/UBC  J. Cruz / A. Roca / JBCVC F. Oliva / A. Marrero / JBCVC /Ojeda 189/UBC Molero 1992 (Fernandez Casas 13699) ORT # 36336 F. Oliva / A. Marrero / JBCVC /Ojeda 182/UBC F. Oliva / A. Marrero / JBCVC /Ojeda 181/UBC A. Roca / B. Navarro / A. Marrerro / JBCVC Ojeda 210  Ojeda 213/UBC Ojeda 228/UBC Ojeda 207/UBC Ojeda 243/UBC A. Santos F. Oliva / J. Navarro / J. Naranjo / B.  105  Taxon Lotus glaucus Sol. Lotus hillebrandii Christ Lotus holosericeus Webb & Berthel. Lotus jacobaeus L.  Lotus jolyi Battand. *Lotus kunkelii (Esteve) Bramwell & D. H. Davis Lotus lancerottensis Webb & Berth. Lotus latifolius Brand Lotus leptophyllus (Lowe) K. Larsen *Lotus macranthus Lowe Lotus maroccanus Ball Lotus mascaensis Burchard  Lotus pseudocreticus Maire, Weiller & Wilczek Lotus purpureus Webb Lotus salvagensis R.P. Murray *Lotus sessilifolius D.C. subsp. villosissimus (Pitard) Sandra & Sokoloff *Lotus sessilifolius D.C. subsp. sessilifolius *Lotus sessilifolius DC. var. pentaphyllus (Link) D. H. Davis Lotus spartioides Webb & Berthel. Lotus tenellus (R. Lowe) Sandral, Santos & D.D. Sokoloff *Lotus sp. nov. 1 Lotus section Rhyncholotus (Monod) D.D. Sokoloff *L. berthelotii Masf. var. berthelotii *Lotus eremiticus A. Santos *L. maculatus Breitf. *Lotus pyranthus P. Perez  Collection information Cultivated JBCVC # 235B/07, Porto Moniz, M DNA Bank # 656, cultivated JBCVC # 42/B, Llano las Chozas, P DNA Bank # 657, cultivated JBCVC # 334/02, Pilancones, GC DNA Bank # 658, cultivated JBCVC # 46/03, Bordeira bei Piorno Fogo CV, 2100 m. Voucher, Province Guelmin, Morocco DNA Bank # 3805, Barranco Jinamar, GC DNA Bank # 3823 Villaverde, Betancuira, F DNA Bank # 1812, Ctra. Porto Novo, Santo Antao, CV Barranco Guayedra, T  Voucher, herbarium or GenBank F. Oliva / A. Marrero / JBCVC Ojeda 187/UBC P. Maya / V. Montelongo / J. Naranjo / R. Febles / JBCVC F. Oliva, / J. Naranjo / J. Navarro / I. Santana / B. Vilches / JBCVC T.Leyens  Voucher, M Voucher, Talouine, Morocco DNA Bank # 659, Cultivated JBCVC # 133/M Valle de Masca, Teno, T Cultivated JAO 468-00  ORT # 36675 S.L. Jury 14471/RNG M. Aleman /JBCVC Ojeda 200/UBC  Cultivated JAO # 130-99 Voucher, SG Las Playas, S from Parador, H  ORT # 36670 ORT # 35118 Ojeda 196/UBC  Poris de Abona, T Punta Llana, G Playa Pocito, Mazo, P San Juan-Guia de Isora, T  Ojeda 225/UBC Ojeda 208/UBC Ojeda 205/UBC  DNA Bank # 662, cultivated JBCVC # 337/02, Chira-Pinar Santiago, GC Arachico, Ermita San Roque, T  F. Oliva / J. Navarro / J. Naranjo / B. Navarro / I. Santana / B.Vilches /JBCVC Ojeda 246/UBC  South Roque dos hermanos, Anaga, T  Ojeda 193/UBC  Ifonche, T Cultivated UBC, commercial plant DNA Bank # 3838, Garafia, P DNA Bank # 660, Puertito Sauzal, T DNA Bank # 661 JBCVC 210/99 DNA Bank # 3842  Ojeda 238/UBC JCVC 366-04 Jaén 08/04 JCVC Ojeda 175/UBC -  S.L. Jury & T.M. Upson 20480/RNG F. Oliva / J. Navarro / J. Caujapé / N. Cabrera /JBCVC F. Oliva / J. Navarro /JBCVC A. Marrero / R Almeida / J. Caujapé/ JBCVC Ojeda 170/UBC  Ojeda 231/UBC  106  Table 4.2 Species included in the phylogenetic analysis with the nuclear ribosomal ITS region only. GenBank sequences and new sequences generated in this analysis. I excluded six sequences from GenBank (L. dumetorum AY294294, L. campylocladus AF450196, L. creticus AF450192, L. arinagensis FJ411112, L. loweanus FJ 411117, L. mascaensis FJ411118 ) due to the ambiguities in the sequences and/or suspected misidentification. JAO= Jardín de Aclimatación de la Orotava, JBCVC= Jardín Botánico Canario “Viera y Clavijo”, UBC= University of British Columbia. T= Tenerife, GC= Gran Canaria, G= La Gomera, P=La Palma, H=El Hierro, CV=Cape Verde, M=Madeira  107  Taxon Outgroup Hosackia chihuahuana S. Watson Hosackia gracilis Benth. Lotus japonicus MG20 (Regel) K. Larsen Lotus japonicus Gifu B129 (Regel) K. Larsen Lotus filicaulis Durieu Lotus corniculatus L. Lotus burttii Borsos Ingroup Lotus section Pedrosia (Lowe) Christ. Lotus arborescens Lowe ex Cout. Lotus arenarius Brot.  Lotus arinagensis Brawm.  Lotus argyrodes R.P Murray  Lotus assakensis Brand  Lotus azoricus P. W. Ball Lotus bollei Christ [=L. purpureus sensu Sandral et al] Lotus brunneri Webb in Hooker Lotus callis-viridis Bramwell & D.H. Davis Lotus campylocladus Webb & Berthel. Lotus creticus L.  Lotus dumetorum Webb ex R. P. Murray Lotus emeroides R. P. Murray  Collection Info.  Voucher, herbarium or GenBank  Cultivated UBC # PI 262405 Mexico  Ojeda 79/UBC  Cultivated UBC Cultivated UBC  Ojeda 69/UBC  Cultivated UBC  Ojeda 70/UBC  Cultivated UBC Vancouver, BC Cultivated UBC seeds from Univ. Miyazaki  Ojeda 71/UBC Ojeda 46/UBC Ojeda 72/UBC  Cultivated JBCVC # 164/06, Morro Cove Roche Sao Nicolao CV Cultivated UBC # PI 631956, Kourigba, Morocco Cultivated UBC # PI 631780, Tiznir, Morocco GenBank GenBank DNA bank # 651 cultivated JBCVC # 138/01 Barranco Viejo, Arinaga, GC DNA bank # 652, Cultivated JBCVC # 55/02 Carretera faro Arinaga, GC Playa Arinaga, GC Cultivated JBCVC # 5485B/UPM/07 (Banco Germoplasma UPM, Dr. Gómez Campo) Ponta Pargo, M Voucher, Punta San Lorenzo, M Voucher, Tarfaya-Tan Tan Sahara, Africa GenBank GenBank DNA Bank # 1084 Tiznir, Aglou Plage, Morocco DNA Bank # 1085 North Agadir, Tamri, Morocco Genbank , Cult. JAO # 161-00, Azores  F. Oliva / A. Marrero / JBCVC / Ojeda 180/ UBC Ojeda 78/UBC AF450193 (Allan et al., 2003) AF218528 (Allan et al, 2004)  Cultivated JBCVC # 163/06, Baia das Gatas on Monte Verde, Sao Vicente, CV Cultivated JBCVC # 514B/07, Mocete Negro (between Stª María and Pedro Lume) Sal, CV DNA Bank # 654, cultivated JBCVC, # 369/04 Andén Verde, GC Road to Cañada Teide, T Arona-Ifonche, T Cultivated JBCVC # 64/05 (# 339/97) Cabo Pino. Cult # PI 308978, Israel Cult # PI 505409, Spain GenBank Mirador Jardina, Mercedes, Anaga, T Teno Alto, Teno, T Inchora, G Epina, G Genbank  J. Cruz / A. Roca / JBCVC J. Naranjo / J. Navarro / F. Oliva / JBCVC Ojeda 202/UBC F. Oliva / A. Marrero / JBCVC /Ojeda 189/UBC ORT # 37806 Molero 1992 (Fernandez Casas 13699) DQ160277 (Degtjareva et al., 2006) AF450204 (Allan et al., 2003) A. Marrero JBCVC A. Marrero JBCVC AY294293 (Allan et al., 2004), ORT # 36336 F. Oliva / A. Marrero / JBCVC /Ojeda 182/UBC F. Oliva / A. Marrero / JBCVC /Ojeda 181/UBC A. Roca / B. Navarro / A. Marrerro / JBCVC Ojeda 169/UBC Ojeda 206/UBC Ojeda 210/UBC B. Navarro /JBCVC / Ojeda 188/UBC Ojeda 241/UBC Ojeda 242/UBC DQ160279 (Degtjareva et al., 2006) Ojeda 213/UBC Ojeda 228/UBC Ojeda 207/UBC Ojeda 209/UBC AY294295 (Allan et al., 2004)  108  Taxon  Collection Info.  Voucher, herbarium or GenBank  Lotus eriosolen (Maire) Mader & Podlech  Cult. UBC # PI 631959, Ouarzate, Morocco Cult. UBC # PI 631784, Tiznir, Morocco GenBank (labeled as L. maroccanus) GenBank Fuerteventura GenBank DNA Bank # 655, cultivated JBCVC # 330/02  Ojeda 244/UBC Ojeda 243/UBC AF450181 (Allan et al., 2003) DQ160281 (Degtjareva et al., 2006) A. Santos AY294296 (Allan et al., 2004) F. Oliva / J. Navarro / J. Naranjo / B.  Cultivated JBCVC # 235B/07, Porto Moniz, M  F. Oliva / A. Marrero / JBCVC Ojeda 187/UBC Ojeda 233/UBC P. Maya / V. Montelongo / J. Naranjo / R. Febles / JBCVC Ojeda 198/UBC Ojeda 232/UBC AY294298 (Allan et al., 2004) F. Oliva, / J. Naranjo / J. Navarro / I. Santana / B. Vilches / JBCVC T.Leyens  L. erythrorhyzus Bolle L. genistoides Webb. (nom. nudum) Lotus glaucus Sol.  Lotus hillebrandii Christ  Lotus holosericeus Webb & Berthel. Lotus jacobaeus L.  Lotus jolyi Battand.  Lotus kunkelii (Esteve) Bramwell & D. H. Davis  Cult. JAO # 19-05 DNA Bank # 656, cultivated JBCVC # 42/B, Llano las Chozas, P Mirador Isora, H Todoque-Playa Naos, P Genbank DNA Bank # 657, cultivated JBCVC # 334/02, Pilancones, GC DNA Bank # 658, cultivated JBCVC # 46/03, Bordeira bei Piorno Fogo CV, 2100 m. DNA Bank # 2089 Ribeira Monte espia Fogo CV DNA Bank # 2181 Fogo, CV DNA Bank # 2126 Ribeira-Campana Fogo, CV Genbak Voucher, Province Tan Tan, Morocco Voucher, Province Guelmin, Morocco GenBank DNA Bank # 3804, Barranco Jinamar, GC DNA Bank # 3805, Barranco Jinamar, GC  DNA Bank # 3823, Villaverde, La Matilla, DNA Bank # 3825 Corralejo, F Voucher, L GenBank DNA Bank # 1812, Ctra. Porto Novo, Santo Antao, CV Genbank Barranco Guayedra, T Voucher, Pico Branco, Porto Santo, M Voucher, M Voucher, Talouine, Morocco Voucher, Marrakech, Morocco DNA Bank # 659, Cultivated JBCVC # 133/M DNA bank # 3844, Masca, T Valle de Masca, Teno, T Punta Teno, Teno, T GenBank Cult. JAO 468-00 Genoank  S.L. Jury & T.M. Upson 20503/RNG S.L. Jury & T.M. Upson 20480/RNG DQ166240 (Degtjareva et al., 2006) F. Oliva / J. Navarro / J. Caujapé / N. Cabrera /JBCVC M. Aleman / JBCVC/ Ojeda 176/UBC F. Oliva / J. Navarro /JBCVC ORT # 36458 AY294300 (Allan et al., 2004) A. Marrero / R Almeida / J. Caujapé/ JBCVC AY294301 (Allan et al., 2004) Ojeda 170/UBC ORT # 33596 ORT # 36675 S.L. Jury 14471/RNG Fernandez Casas 13737/RNG M. Aleman /JBCVC Ojeda 200/UBC Ojeda 230/UBC AY294302 (Allan et al., 2004) Ojeda 231/UBC DQ160284 (Degtjareva et al., 2006)  Cult. JAO # 130-99 Genbank Voucher, Salvage Grande, SG  ORT # 36670 AY294303 (Allan et al., 2004) ORT # 35118  Cultivated JBCVC # 217/07 ( # 497/99) Lotus lancerottensis Webb & Berth.  Lotus latifolius Brand Lotus leptophyllus (Lowe) K. Larsen Lotus macranthus Lowe Lotus maroccanus Ball Lotus mascaensis Burchard  Lotus pseudocreticus Maire, Weiller & Wilczek Lotus purpureus Webb Lotus salvagensis R.P. Murray  A. Marrero / R. Almeida / J. Caujapé Marrero et al JCVC Marrero et al JCVC AY294299 (Allan et al., 2004)  109  Taxon Lotus sessilifolius D.C. subsp. villosissimus (Pitard) Sandral & Sokoloff  Collection Info. Las Playas, S from Parador, H  Voucher, herbarium or GenBank Ojeda 196/UBC  Lotus sessilifolius D.C. subsp. sessilifolius  Poris de Abona, T Punta Llana, G Playa Pocito, Mazo, P San Juan-Guia de Isora, T  Ojeda 225/UBC Ojeda 208/UBC Ojeda 205/UBC  Tamadaba, GC  F Oliva/J. Caujapé /R. Jaén /JBCVC Ojeda 217/UBC F Oliva/J. Caujapé /R. Jaén /JBCVC Ojeda 216/UBC AY294304 (Allan et al., 2004) F. Oliva / J. Navarro / J. Naranjo / B. Navarro / I. Santana / B.Vilches /JBCVC F Oliva/J. Caujapé /R. Jaén /JBCVC, Ojeda 211/UBC F Oliva/J. Caujapé /R. Jaén /JBCVC, Ojeda 191/UBC AY294305 (Allan et al., 2004) AY294297 (Allan et al., 2004) Ojeda 446/UBC Ojeda 215/UBC A. Santos Ojeda 193/UBC Ojeda 194/UBC Ojeda 203/UBC Ojeda 167/UBC  Lotus sessilifolius DC. var. pentaphyllus (Link) D. H. Davis Lotus spartioides Webb & Berthel.  Pinar Pajonales, GC GenBank DNA Bank # 662, cultivated JBCVC # 337/02, Chira-Pinar Santiago, GC Presa las Niñas, GC Llanos de la Pez, GC  Lotus tenellus (R. Lowe) Sandral, Santos & D.D. Sokoloff Lotus sp. nov. 1  Lotus sp. nov. 2 Lotus sp. nov. 3 Lotus section Rhyncholotus (Monod) D.D. Sokoloff L. berthelotii Masf.  Lotus eremiticus A. Santos L. maculatus Breitf.  Lotus pyranthus P. Perez  GenBank, T GenBank, T (labeled as L. glaucus) Arachico, Ermita San Roque, T Playa de los Roques, Tagana, Anaga, T Punta Hidalgo, Anaga, T South Roque dos hermanos, Anaga, T Teno Alto, Teno, T Cortijo de San Ignacio, GC Punta Góngora, GC  DNA Bank # 3832, Arico, T Ifonche, T Cultivated UBC, commercial plant GenBank Cultivated JAO 430-95 DNA Bank # 3838, Garafia, P GenBank DNA Bank # 660, cultivated JBCVC # 43/99, ex horto Puertito Sauzal, T Cultivated UBC, commercial plant DNA Bank # 3840 Sauzal, T GenBank DNA Bank # 661, cultivated JBCVC, # 210/99 DNA Bank # 3842, cultivated Vivero Ceplam, Bco. Cultivated JAO 124/01 GenBank  Ojeda 238/UBC AY204306 (Allan et al., 2004) JBCVC 366-04 AY294307 (Allan et al., 2004) R. Almeida / JBCVC Ojeda 239 AY294308 (Allan et al., 2004) J. Cruz/ JBCVC / Ojeda 175/UBC F. Oliva / E. Ojeda Ojeda 226/UBC AY294309 (Allan et al., 2004)  110  Table 4.3 Nuclear and plastid regions used in the three data sets analyzed in this study.  Locus Nuclear ITS  LjCyc1 LjCyc2  LjCyc3 LEGCYC  Plastid matK CYB6  trnHpsbA  Primer sequence (5'--3')  ITS F5 GGAAGGAGAAGTCGTAACAAG ITS R4 TCCTCCGCTTATTGATATGC Cyc1.1F TTCTCCTTCACCATACCC Cyc1.1R TTGGATACATAGGGAAGG LC2.1F TCCCTTTCAGCTCAAGCCCTTACCC LC2.1R GAAGTCATCTCTTGGCGCCTCACC CYC3.2F ACTCCATTAACCCTTTCCS CYC3.1R CCTGCTTCCTTATTAGGGATTGC LEGCYCF TCAGGGSYTGAGGGACCG LEGCYCR TCCCTTGCTCTTGCTCTTGC  matKXf TAATTTACGATCAATTCATTC matk 3.2R CTTCCTCTGTAAAGAATTC CYB6F CTTTTTGTTTTGAGCCGTACGAGATGA CYB6R AAGTCATAGCAAAACCCGTCGCTACT trnH GGCGCATGGTGGATTCACAAATC psbA GTTATGCATGAACGTAATGCTC  Region  Size (bp)  Reference  Intergenic ribosomal region Cycloidea homologue Cycloidea homologue  660  White et al., 1990  935  This study  932  Cronk, unpubl.  Cycloidea homologue  982  This study  Region between the TCP and R domains in Legumes  Citerne et al., 2003  Maturase K  980  Kress et al., 2005  Cytochrom e B6  183  Choi et al., 2006  Intergenic region  367  Kress et al., 2005  111  Table 4.4 Information of the four nuclear and plastid regions used in the phylogenetic reconstruction of Pedrosia s.l. with the ITS only including GenBank sequencesa, excluding ITS sequences from GenBankb and with the ITS data set combined with the plastid regionsc. *Using maximum parsimony.  a  ITS ITSb  CYB6  Plastid combined  All four regions  ITS  125  100  54  54  54  54  54  54  621  621  626  385  898  183  1465  2093  2  2  2  2  0  0  1  2  390  403  446  265  848  175  1288  1735  105  103  100  77  23  4  107  209  23  22  16  7  8  1  17  34  303  285  239  135  56  8  37  451  ITS Number of samples Aligned length (bp) Number of indels (ingroup) Number of constant sites Number of parsimony informative sites Number of parsimony informative sites excluding outgroup Number of trees recovered*  c  Plastid trnH-psbA matK  112  Table 4.5 Gene regions used with the 21 sample data set used to identify the closest relative species of the four rhyncholotus species within clade B. Variability of each region when analyzed separate and in combination.  Number of samples Aligned length (bp) Number of indels (ingroup) Number of constant sites Number of parsimony informative sites Number of parsimony informative sites excluding outgroup Number of trees recovered  ITS 21  Nuclear regions LjCYC1 LjCYC2 LjCYC3 21 21 21  Plastid regions trnH-psbA matK 21 21  All combined 21  615  879  851  880  340  898  4461  0  11  4  2  2  0  19  55  826  783  825  326  895  4193  4  29  27  12  6  3  80  2  24  26  7  6  3  67  1  30  24  60  18  20  374  Table 4.6 Date of origin (MRCA) in Ma of various clades based on two data sets. Values obtained for each clade when rhyncholotus constrained or unconstrained to be monophyletic.  Clade [L. sessilifolius + rhyncholotus] [Rhyncholotus] [L. eremiticus/L. pyranthus] [L. berthelotii/L. maculatus]  4-gene chronogram  6-gene chronogram  Constrained  Unconstrained  Constrained  Unconstrained  2.20 1.20(0.32-1.70) -  2.08 -  2.70((0.62-4.17) 1.72(0.29-3.3) 0.58(0.06-1.81) 1.45(0.06-1.80)  2.09(0.62-4.13) 0.67(0.06-1.90) 1.11(0.07-2.21)  113  Figure 4.1 Geographical distribution of Lotus section Pedrosia and Rhyncholotus in the Macaronesian region, including five Atlantic volcanic archipelagos (Madeira, Azores, the Salvage islands, the Canary Islands and Cape Verde islands), Europe and Africa. Each archipelago with the No. of species/ No. of endemic species. The age of current above-sea level for each island according to Carracedo et al. (2002). The phytogeographic region of Macaronesia is indicated in dashed lines, including a portion of Africa mainland denominated as the “Macaronesian enclave” (Kim et al. 2008).  Azores (9 islands, 4-8 Ma) 2/1 L. azoricus L. creticus  L. creticus (Mediterranean) L. arenarius (Spain, Portugal Gibraltar, Morocco)  Madeira (3 islands, 5 Ma) 5/4 L. argyrodes L. macranthus L. loweanus L. glaucus L. lancerottensis  Canary Islands (7 islands, 1.12-21 Ma)  Salvage Islands (2 islands, 10 Ma) 1/1 L. salvagensis  23/23  Morocco 8/4 L. arenarius L. creticus L. assakensis L. maroccanus L. eriosolen L. pseudocreticus L. jolyi (Mauritania, Chad, Algeria W. Sahara) L. chazalei (W. Sahara, S-W Morocco)  Cape Verde (10 islands, 6-20 Ma) 6/6  L. chazalei (Mauritania) L. tibesticus (Chad)  L. arborescens L. bollei L. brunneri L. purpureus L. jacobaeus L. latifolius  114  Figure 4.2 Geographical distribution of Lotus sections Pedrosia and Rhyncholotus in the Canary Island archipelago, with the oldest ages of the subaerial volcanism of each island according to Carracedo et al. (2002). Species in black are bird-pollinated species from the rhyncholotus group. No. of species/No. of endemic species on each island.  Lanzarote 15.5 Ma 1/0  La Palma  Fuerteventura  1.77 Ma 4/2  20.6 Ma 2/1  L. eremiticus L. pyranthus L. hillebrandii L. seesilifolius subsp . sessilifolius  Tenerife 11.6 Ma  L. lancerottensis  Gran Canaria 14.5 Ma  8/7  9/8  La Gomera 12 Ma 2/1  El Hierro 1.12 Ma 2/1  L. sessilifolius subsp . sessilifolius L. emeroides  L. sessilifolius subsp . villosisimus L. hillebrandii  L. dumetorum L. tenellus L. sessilifolius subsp . sessilifolius L. mascaensis L. campylocladus L. berthelotii L. maculatus Lotus sp . nov. 1. Teno -Anaga  L. lancerottensis L. erytrhorhizus  L. kunkelii L. arinagensis L. spartioides L. callis - viridis L. leptophyllus L. holosericeus L. genistioides Lotus sp . nov. 2 Punta Hidalgo Lotus sp . nov. 3 Cortijo San Ignacio  115  Figure 4.3 Strict consensus tree of the phylogenetic relationships using parsimony within Pedrosia s.l. using ITS DNA sequences. The four clades recovered are labeled (clade A-D). * indicate species groups previously recognized in Lotus according to Sandral et al (2006). Values above branches represent bootstrap values/Bayesian support.  116  Hosackia chihuahuana Hosackia gracilis L japonicus Gifu B129 L japonicus MG20 L burtii L filicaulis L corniculatus  62/99  63/87 62/98  -/71  Clade A  58/71  Outgroup  L. arborescens Sao Nicolao L. bollei Sao Vicente L. brunneri Sal L. jacobaeus Fogo L. jacobaeus Fogo L. jacobaeus AY294299 L. jacobaeus Sao Jorge L jacobaeus Fogo L. purpureus L. latifolius Santo Antao L. purpureus AY294303  Taxonomic group  L. purpureus group  L. eriosolen Quarzate, Morocco L. eriosolen Tiznit, Morocco L. eriosolen DQ160281 L. eriosolen AF450181 L. arenarius Kourigba, Morocco L. arenarius Tiznit, Morocco L. arenarius AF450193 L. arenarius AF218528 L. maroccanus Talouine, Morocco L. maroccanus Marackech, Morocco L. arinagensis Arinaga, Gran Canaria L. arinagensis Arinaga, Gran Canaria L. arinagensis Arinaga, Gran Canaria L. emeroides Inchora, La Gomera L. emeroides Epina, La Gomera L. emeroides AY294295 L. kunkelii Jinamar, Gran Canaria L. kunkelii Jinamar, Gran Canaria L. kunkelii Jinamar, Gran Canaria L. mascaensis Masca, Tenerife L. mascaensis AY294302 L. mascaensis Masca, Tenerife L. mascaensis Punta Teno, Tenerife L. mascaensis Masca, Tenerife L. sessilifolius subsp. sessilifoliusLa Gomera L. sessilifolius subsp. sessilifolius Poris de Abona, Tenerife L. sessilifolius var. pentaphyllus San Juan Guia-Isora, Tenerife L. sessilifolius subsp. sessilifolius La Palma L. sessilifolius subsp. villossisimus Hierro  L. arenarius group*  54/94  Major distribution  Cape Verde  Africa, Spain  L. sessilifolius group*  60/96  L. berthelotii L. berthelotii L. berthelotii L. berthelotii L. eremiticus L. eremiticus L. eremiticus L. maculatus L. maculatus L. maculatus L. maculatus L. pyranthus L. pyranthus L. pyranthus L. pyranthus  Clade B -/83  Cult. UBC AY294306 Arico, Tenerife Ifonche, Tenerife AY294307 Garafia, La Palma Garafia, La Palma El Sauzal, Tenerife Cult. UBC El Sauzal, Tenerife AY294308 AY294309 Cult. JCVC Cult. JCVC Cult. JCVC  Rhyncholotus  L. azoricus AY294293 L. argyrodes Madeira L. argyrodes Madeira L. macranthus Madeira L. macranthus Madeira Lotus sp. nov. 1 Playa Los Roques, Anaga, Tenerife Lotus sp nov 1 Punta Hidalgo, Anaga, Tenerife Lotus sp. nov .1 Roque dos Hermanos Anaga, Tenerife Lotus sp. nov .1 Teno Alto, Tenerife  -/51  65/-  59/90  Clade C 82/94  61/- 84  L. callis -viridis Anden Verde, Gran Canaria L. callis -viridis Anden Verde, Gran Canaria L. campylocladus Carretera Cañadas, Tenerife L. campylocladus Arona-Ifonche, Tenerife L. genistioides Cañadon Sombrio, Gran Canaria L. hillebrandii Cumbrecita, La Palma L. hillebrandii AY294298 L. hillebrandii Todoque, La Palma L. hillebrandii Mirador, Hierro L. holosericeus Pilancones, Gran Canaria L. spartioides Tamadaba, Gran Canaria L. spartoides Pinar Pajonales L. spartioides AY294304 L. spartioides Pinar Santiago, Gran Canaria L. spartioides Presa Las Ninas, Gran Canaria L. spartioides Llanos de la Pez, Gran Canaria L. dumetorum Anaga, Tenerife L. dumetorum Carretra Baraderos Anaga, Tenerife L. tenellus AY294305  63/-  L. argyrodes Madeira, group Azores  L. campylocladus  L. dumetorum Teno, Tenerife L. dumetorum Chinamada, Teno, Tenerife Lotus sp. nov 3 Punta Gongora, Gran Canaria Lotus sp. nov. 2 Cortijo San Ignacio, Gran Canaria L. lancerottensis AY294300 L. lancerottensis Fuerteventura L lancerottensis Lanzarote L lancerottensis Lanzarote L. erythrorhizus AY294296 L. erythrorhizus Fuerteventura L. leptophyllus AY294301 L. leptophyllus Gran Canaria L. tenellus AY294297 L. tenellus Arachico, Tenerife L. glaucus Madeira L. glaucus Madeira L. salvagensis Salvage Islands L. assakensis DQ160277 L. assakensis AF450204 L assakensis Morocco L assakensis N. Agadir, Tamri, Morocco L assakensis Tiznit, Aglou Plage, Morocco L. creticus DQ160279 L. creticus JCVC 379/97 L. creticus Israel L. creticus Spain L. pseudocreticus DQ160284 L. pseudocreticus JAO 468-00  Clade D  Canary Islands  L. jolyi Guelmin, Morocco L. jolyi Tan Tan Morocco L. jolyi DQ166240  group* Canary Islands  L. glaucus group* Madeira Salvages L. assakensis group*  Africa, Mediterranean  L. jolyi group*  117  Figure 4.4 Strict consensus tree based on a combined analysis of three plastid regions (trnH-psbA, matK and CYB6) using maximum parsimony. Values above branches indicate bootstrap values.  88 66  64 58  78  91  91  98  63  H. chihuahuana H. gracilis L. japonicus MG20 L. japonicus GIFU B129 L. filicaulis L. burtii L. corniculatus L. arborescens L. brunneri L. jacobaeus L. bollei L. latifolius L. eriosolen L. joyli L. purpureus L. arenarius L. arinagensis L. assakensis L. creticus L. pseudocreticus L. erythrorhizus L. glaucus L. kunkelii L. lancerottensis L. genistioides L. argyrodes L. azoricus L. macranthus L. callis-viridis L. mascaensis L. sessilifolius Abona, Tenerife L. sessilifolius La Gomera L. sessilifolius La Palma Lotus sp. nov. 1 Tenerife L. berthelotii L. berthelotii Bird-pollinated L. eremiticus L. maculatus Rhyncholotus L. pyranthus L. pyranthus L. campylocladus L. dumetorum Anaga L. dumetorum Teno L. emeroides L. sessilifolius El Hierro L. sessilifolius var. pentaphyllus L. hillebrandii L. holosericeus L. leptophyllus L. maroccanus L. mascaensis L. salvagensis L. spartioides L. tenellus  118  Figure 4.5 Majority tree recovered from a combined analysis of one nuclear (ITS) and three plastid gene regions (matK, trnH-psbA and CYB6) using maximum parsimony. Branches with an arrow indicate clades not observed in the strict consensus. Values above branches represent bootstrap from MP/posterior probabilities support.  119  84/99 98/100 100/99  75/78  95/100 59/-/86 Clade A  75/99 77/99  100/100  -/93  100/90  Clade B -/97 95/100  100  66/89  52/99 Clade C  Clade D  Hosackia chihuahuana H. gracilis L. japonicus MG20 Outgroup L. japonicus GIFU B129 L. burtii L. filicaulis L. corniculatus L. arborescens L. brunneri L. jacobaeus Cape Verde L. bollei L. latifolius L. purpureus L. eriosolen L. arenarius Africa L. maroccanus L. arinagensis Gran Canaria L. kunkelii Gran Canaria L. emeroides La Gomera L. mascaensisTenerife L. sessilifoliussubsp. sessilifoliusTenerife L. sessilifoliussubsp. sessilifoliusLa Gomera L. sessilifoliussubsp. sessilifoliusLa Palma Lotus sp. nov. 1 Tenerife L. berthelotii Tenerife L. berthelotii Tenerife L. eremiticus La Palma Bird-pollinated L. pyranthus La Palma Rhyncholotus L. maculatus Tenerife L. pyranthus La Palma L. sessilifolius subsp. villossisimus El Hierro L. sessilifoliusvar. pentaphyllus Tenerife L. mascaensis Tenerfie L. argyrodes Madeira L. azoricus Azores L. macranthus Madeira L. assakensis Africa L. callis-viridis Gran Canaria L. genistioides Gran Canaria L. holosericeus Gran Canaria L. campylocladus Tenerife L. hillebrandii La Palma L. spartioides Gran Canaria L.dumetorum Tenerife L. dumetorum Tenerife L. creticus Africa L. pseudocreticus Africa L. erythrorhizus Fuerteventura L. glaucus Madeira L. lancerottensis Lanzarote L. leptophyllus Gran Canaria L. salvagensis Salvage Islands L. tenellus Tenerife L. jolyi Africa  120  Figure 4.6A. Maximum parsimony strict consensus tree based on four nuclear regions (ITS, LjCYC 1, 2 and 3) and two plastid regions (matK and trnH-psbA) using maximum parsimony. Values above branches indicate bootstrap values/ Bayesian support. Values below 50 are not indicated. Arrows indicate clades recovered in MP, Bayesian and ML analyses using this data set. L. japonicus GIFU B129 L. arenarius, Africa 75/98 L. argyrodes, Madeira 100/99 L. macranthus, Madeira L. azoricus Azores Lotus sp. nov. 1, Tenerife 56/69  L. kunkelii, Gran Canaria  L. arinagensis, Gran Canaria L. mascaensis Valle de Masca, Tenerife 76/54 L. mascaensis Valle de Masca, Tenerife L. sessilifolius subsp. villosissimus, El Hierro 5 changes L. sessilifolius subsp. sessilifolius, La Palma 76/66  L. sessilifolius subsp. sessilifolius, Tenerife 73/100 L. eremiticus, La Palma L. pyranthus, La Palma -/99  L. pyranthus, La Palma  L. sessilifolius subsp. sessilifolius, La Gomera 65/100 L. berthelotii Ifonche, Tenerife L. berthelotii, Tenerife -/99 L. maculatus, Puertito Sauzal, Tenerife L. sessilifolius var. pentaphyllus, Tenerife L. emeroides Inchora, La Gomera  121  Figure 4.6B. Maximum likelihood tree recovered using a data set of six genes and 21 samples with Garli.  L. japonicas, GIFU B129 L. arenarius, Africa L. azoricu,s Azores L. argyrodes, Madeira L. macranthus, Madeira L. arinagensis, Gran Canaria L. kunkeli, Gran Canaria L. emeroides, La Gomera  L. sessilifolius group + rhyncholotus clade  L. mascaensis, Tenerife Lotus sp. nov.1, Tenerife L. mascaensis L. sessilifolius var. pentaphyllus, Tenerife  L. berthelotii/L. maculatus subclade  L. maculatus, Tenerife L. berthelotii Ifonche, Tenerife L. berthelotii, Tenerife L. sessilifolius subsp. sessilifolius, La Palma L. sessilifolius subsp. villossisimus, El Hierro L. sessilifolius subsp. sessilifolius, La Gomera  L. pyranthus/L. eremitcus subclade  L. sessilifolius subsp. sessilifolius, Tenerife L. pyranthus, La Palma L. pyranthus, La Palma L. eremiticus, La Palma  122  Figure 4.7 Flower and leaflet morphology in L. argyrodes (A-C), L. mascaensis (D-F) and L. sessilifolius var. pentaphyllus (G-I).  A  B  C  D  E  F  G  H  I  123  Figure 4.8 Flower and leaflet morphology in the four bird-pollinated species. Both groups are differentiated by minor morphological features including the number of flowers per inflorescence, the position and orientation of the dorsal petal and the position of the lateral petal. In the species from La Palma the lateral petal is fused at the tip and covers the tip of the ventral petal. In contrast, species from Tenerife have a ventral petal more exposed.  Tenerife  Lotus berthelotii  Lotus maculatus  La Palma  Lotus eremiticus  Lotus pyranthus  124  Figure 4.9 Chronogram obtained for the evolution of bird pollination in Macaronesian Lotus under a Bayesian relaxed clock uncorrelated clock model using Beast and applied to the combined data set of 52 samples and using a data set of four gene regions (ITS, matK, trnH-psbA and CYB6). Upper limits of the ages of La Palma (1.77 Ma), El Hierro (1.12 Ma), and Fuerteventura (21 Ma) were used as calibration points (black circles). Bird-pollinated species are shown in red branches. Ages estimates with their 95% credibility intervals are shown on nodes. Values on grey squares represent bootstrap values from MP/posterior probabilities inferred from the Bayesian inference. Major geological events between the Miocene and Pleistocene are indicated with arrows at the bottom. Clades named after the groups recovered with ITS and the four gene data analyses.  125  12.55(0.82-13.88) 1.57(0.03-1.95)  75/84 100/99  1.57(0.16-3.90)  6.74(0.37-7.52) 100/98  0.36(0-0.69)  6.1(0.50-7.29) 96/92  Clade A+D -/81  16.32(1.55-20.68)  1.33(0.04-2.4)  4.78(0.39-5.19)  9.24(0.76-9.75)  -/75  100/75  4.55(0.11-4.79)  0.37 (0-0.95) 0.65 (0.08-2.31)  MRCA L. sessilifolius + rhyncholotus  0.7 (0-0.99) 2.20 1.2 (0.32-1.76) 0.1(0-0.73)  11.97 (1.2-12.44) 1.74  100/79 3.71(0.39-4.81)  Clade B  -/89 5.89(0.66-6.92)  1.74  98/89 2.51(0.07-3.13) 100/95  0.19(0.0-1.66) 0.94(0.03-2.20)  8.45(0.88-9.44) 98/82  3.21(0.13-3.06)  Clade C  0.9(0-1.18)  4.34(0.58-6.38) 2.46(0.05-2.12)  98/91 2.96(0.21-3.90)  Tenerife Gran Canaria (Roque Conde)  Tenerife (Teno and Anaga)  La Gomera  Tenerife (volcanic activity)  Miocene 17.5  15.0  12.5  10.0  La Palma El Hierro  Pleistocene  Pliocene 7.5  5.0  L. corniculatus L. filicaulis L. japonicus MG20 Outgroup L. japonicus GIFU B129 L. burtii L. eriosolen Africa L. bollei L. jacobaeus L. brunneri L. arborescens Cape Verde L. latifolius L. purpureus L. jolyi Africa L. arenarius L. maroccanus L. sessilifolius subsp. sessilifolius La Gomera L. sessilifolius var. pentaphyllus Tenerife L. arinagensis Gran Canaria L. kunkelii Gran Canaria L. pyranthus La Palma L. maculatus Tenerife L. pyranthus La Palma L. berthelotii Tenerife L. eremiticus La Palma L. berthelotii Tenerife L. sessilifolius subsp. villosissimus El Hierro L. sessilifolius subsp. sessilifolius La Palma Lotus sp. nov. 1 Tenerife L. mascaensis Tenerife L. mascaensis Tenerife L. emeroides La Gomera L. mascaensis Tenerife L. azoricus L. argyrodes Madeira, Azores L. macranthus L.genistioides Gran Canaria L.dumetorum Anaga L. callis-viridis Gran Canaria L. glaucus Madeira L. creticus Africa L. assakensis Africa L. lancerottensis Lanzarote L. hillebrandii La Palma L. spartioides Gran Canaria L. salvagensis Salvage islands L. erythrorhizus Fuerteventura L. pseudocreticus Africa L. leptophyllus Gran Canaria L. tenellus Tenerife L. holosericeus Gran Canaria L. campylocladus Tenerife L. dumetorum Teno, Tenerife  2.5  0.0  126  Figure 3.10 Chronogram of the rhyncholotus group using a six gene data set and analyzed with Beast using El Hierro (1.12 Ma), La Palma (1.77 Ma) and the age of Gran Canaria (14.5 Ma) as upper age estimates (black circles). Node ages are indicated above branches with 95% HDP intervals. MRCA L. sessilifolius + rhyncholotus  L. sessilifolius var. pentaphyllus L. sessilifolius Tenerife 1.07(0.5-2.32)  L. sessilifolius La Palma  0.6168 0.7144  2.09(0.62-4.13) 1.23  1.71  MRCA L. pyranthus/ L. eremiticus  2.80(0.7-5.69)  L. sessilifolius El Hierro L. sessilifolius La Gomera  L. eremiticus La Palma 0.52(0.007-0.84) L. pyranthus La Palma 0.67(0.06-1.90) L. pyranthus La Palma  1.11(0.076-2.21  MRCA L. berthelotii/ L. maculatus  L. berthelotii Tenerife 0.09 L. berthelotii Ifonche L. maculatus Tenerife L. emeroides La Gomera L. kunkelii Gran Canaria L. arinagensis Gran Canaria  2.68  Lotus sp. nov. 1 Tenerife L. mascaensis Tenerife L. mascaensis Tenerife L. azoricus Azores L. argyrodes Madeira L. macranthus Madeira L. arenarius Africa L. japonicus 1.0 4.0  3.5  3.0  2.5  2.0  1.5  1.0  0.5  0.0  127  4.6 Bibliography Allan, G.J., Francisco-Ortega, J., Santos-Guerra, A., Boerner, E., Zimmer, E., 2004. Molecular phylogenetic evidence for the geographic origin and classification of Canary Island Lotus (Fabaceae: Loteae). Molecular Phylogenetics and Evolution 32, 123-138. Allan, G.J., Zimmer, E.A., Wagner, W.L., Sokoloff, D.D., 2003. Molecular phylogenetic analyses of tribe Loteae (Leguminosae): implications for classification and biogeography. In: Klitgaard, B.B., Bruneau, A. (Eds.), Advances in Legume systematics. Part 10. Higher level systematics. Royal Botanical Gardens, Kew, pp. 371-393. Ancochea, E., Hernan, F., Cendrero, A., Cantagrel, J.M., Fuster, J.M., Ibarrola, E., Coello, J., 1994. Constructive and destructive episodes in the building of a young oceanic island, La Palma, Canary Island, and genesis of the Caldera de Taburiente. Journal of Volcanology and Geothermal Research 60, 243-262. Anderson, C.L., Channing, A., Zamuner, A.B., 2009. Life, death and fossilization on Gran Canaria–implications for Macaronesian biogeography and molecular dating. Journal of Biogeography 36, 2189-2201. Bañares, A., Blanca, G., Güemes, J., Moreno, J.C., Ortiz, S., 2004. Atlas y Libro Rojo De La Flora Vascular Amenazada de España. Dirección General de Conservación de la Naturaleza, Madrid. Barker, N.P., Weston, P.H., Rutschmann, F., Suaquet, H., 2007. Molecular dating of the "Gondwanan" plant family Proteaceae is only partially congruent with the timing of the break-up of Gondwana. Journal of Biogeography 34, 2012-2027. Bentham, G., 1865. Leguminosae. In: Bentham, G., Hooker, J.D. (Eds.), Genera Plantarum I, pp. 434-600.  128  Bohle, U.R., Hilger, H.H., Martin, W.F., 1996. Island colonization and evolution of the insular woody habit in Echium L. (Boraginaceae). Proceedings of the Natural Academy of Sciences of the United States of America 93, 11740-11745. Bramwell, D., Bramwell, Z., 2001. Flores silvestres de las Islas canarias. Editorial Rueda, Madrid. Brand, A., 1898. Monographie der Gattung Lotus. Botanische Jahrbücher für Systematik 25, 166232. Bräuchler, C., Meimberg, H., Heubl, G., 2004. Molecular phylogeny of the genera Digitalis L. and Isoplexis (Lindley) Loudon (Veronicaceae) based on ITS- and trnL-F sequences. Plant Systematics and Evolution 248, 111-128. Caceres, K.M., Santos-Guerra, A., Coello, R.M., 2004a. Lotus eremiticus A. Santos. In: Bañares, A.B., Blanca, G., Guemes, J., Moreno, J.C., Ortiz, S. (Eds.), Atlas y libro rojo de la flora vascular amenazada de España. Dirección General para la Biodiversidad, Publicaciones de la O.A.P.N., Madrid, pp. 392-393. Caceres, K.M., Santos-Guerra, A., Coello, R.M., 2004b. Lotus pyranthus A. Santos. In: Bañares, A.B., Blanca, G., Guemes, J., Moreno, J.C., Ortiz, S. (Eds.), Atlas y libro rojo de la flora vascular amenazada de España. Dirección General para la Biodiversidad, Publicaciones del O.A.P.N., Madrid, pp. 392-393. Carracedo, J.C., 1994. The Canary Islands: an example of structural control of the growth of large-oceanic islands volcanoes. Journal of Volcanology and Geothermal Research 60, 225241. Carracedo, J.C., Pérez, F.J., Ancochea, E., J., M., Hernán, F., C.R., C., Casillas, R., Rodriguez, E., Ahijado, A., 2002. Cenozoic volcanism II: The Canary Islands. In: Gibbons, W.,  129  Moreno, T. (Eds.), The Geology of Spain The Geological Society Publishing House. , London, pp. 439-472. Choi, H.K., Luckow, M.A., Doyle, J.J., Cook, D.R., 2006. Development of nuclear gene-derived molecular markers linked to legume genetic maps. Molecular Genetics and Genomics 276, 56-70. Christ, H., 1888. Spicilegium Canariense. Botanische Jahrbücher für Systematik 9, 113-172. Citerne, H.L., Luo, D., Pennington, R.T., Coen, E., Cronk, Q.C.B., 2003. A phylogenomic investigation of CYCLOIDEA-like TCP genes in the Leguminosae. Plant Physiology 131, 1042-1053. Coello, R.M., 2007. Seguimiento de poblaciones de especies amenazadas L. pyranthus. In: Ambiente, C.d.M. (Ed.). Gobierno de Canarias, La Palma. Cox, S.C., Carranza, S., Brown, R.P., 2010. Divergence times and colonization of the Canary Islands by Gallotia lizards Molecular Phylogenetics and Evolution 56, 747-757. Cronk, Q.C.B., 1992. Relict floras of Atlantic islands: patterns assessed. Biological Journal of the Linnean Society 46, 91-103. Crowell, K.L., 1962. Reduced interspecific competition among the birds of Bermuda. Ecology 43, 75-88. Degtjareva, G.V., Kramina, T.E., Sokoloff, D.D., Samigullin, T.H., Sandral, G., Valiejo-Roman, C.M., 2008. New data on nrITS phylogeny of Lotus (Leguminosae, Loteae). Wulfenia 15, 35-49. Degtjareva, G.V., Kramina, T.E., Sokoloff, D.D., Samigullin, T.H., Valiejo-Roman, C.M., Antonov, A.S., 2006. Phylogeny of the genus Lotus (Leguminosae, Loteae): evidence from nrITS sequences and morphology. Canadian Journal of Botany 84, 813-830.  130  Dietzen, C., 2007. Molecular phylogeography and colonization history of passerine birds of the Atlantic islands (Macaronesia). Combined Faculties for the Natural Sciences and for Mathematics. University of Heidelberg, Germany. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19, 11-15. Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics and dating with confidence. PLoS Biology 4, e88. Drummond, A.J., Rambaut, A., 2007. "BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7, 214. Dupont, Y.L., Hansen, D.M., Rasmussen, J.T., Olesen, J.M., 2004. Evolutionary changes in nectar sugar composition associated with switches between bird and insect pollination: the Canarian bird-flower element revisited. Functional Ecology 18, 670-676. Dupont, Y.L., Skov, C., 2004. Influence of geographical distribution and floral traits on species richness of bees (Hymenoptera: Apoidea) visiting Echium species (Boraginaceae) of the Canary Islands. International Journal of Plant Sciences 165, 377–386. Farris, J.S., 1989. The retention index and homoplasy excess. Systematic Zoology 38, 406-407. Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using bootstrap. Evolution 39, 783-791. Fuertes-Aguilar, J., Ray, J.M., Francisco-Ortega, J., Santos-Guerra, A., Jansen, R.K., 2002. Molecular evidence from chloroplast and nuclear markers for multiple colonizations of Lavatera (Malvaceae) in the Canary Islands. Systematic Botany 27, 74-83. García-Maroto, F., Mañas-Fernández, A., Garrido-Cárdenas, J.A., Alonso, D.L., Guil-Guerrero, J.L., Guzmán, B., Vargas, P., 2009. Delta6-desaturase sequence evidence for explosive  131  Pliocene radiations within the adaptive radiation of Macaronesian Echium (Boraginaceae). Molecular Phylogenetics and Evolution 52, 563-574. Gómez, M.C.M., Coello, R.M., 2004. Lotus maculatus Breift. In: Bañares, A.B., Blanca, G., Guemes, J., Moreno, J.C., Ortiz, S. (Eds.), Atlas y libro rojo de la flora vascular amenazada de España. Dirección General para la Biodiversidad, Publicaciones del O.A.P.N., Madrid, pp. 398-399. Gonzáles, G.R., Pérez de Paz, P.L., Arenciba, M.C.L., Betancourt, J.A.L., 2004. Lotus pyranthus P. Pérez. In: Bañares, A.B., Blanca, G., Guemes, J., Moreno, J.C., Ortiz, S. (Eds.), Atlas y libro rojo de la flora vascular amenazada de España. Dirección General para la Biodiversidad, Publicaciones del O.A.P.N., Madrid, pp. 400-401. Hind, N., 2008. Lotus maculatus. Leguminosae-Papilionoideae. Curtis's Botanical Magazine 25, 146-157. Hodges, S.A., 1997. Rapid radiation due to a key innovation in columbines (Ranunculaceae: Aquilegia). In: Givnish, T.J., Sytsma, K.J. (Eds.), Molecular evolution and adaptive radiation. Cambridge University Press, New York, pp. 391-405. Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17, 754-755. Hughes, M., Möller, M., Edwards, T., Bellstedt, D.U., Villiers, M., 2007. The impact of pollination syndrome and habitat on gene flow: a comparative study of two Streptocarpus (Gesneriaceae) species. American Journal of Botany 94, 1688-1695. Humphries, C.J., 1979. Endemism and evolution in Macaronesia. In: Bramwell, D. (Ed.), Plants and islands. Academic Pres, London, pp. 171-200.  132  Kim, S.C., McGowen, M.R., Lubinsky, P., Barber, J.C., Mort, M.E., Santos-Guerra, A., 2008. Timing and tempo of early and successive adaptive radiations in Macaronesia. PLoS ONE 3, e2139. Kluge, A.G., Farris, J.S., 1969. Quantitative phyletics and the evolution of anuras. Systematic Zoology 18, 1-32. Kramer, A.T., Fant, J.B., Ashley, M.V., 2011. Influences of landscape and pollinators on population genetic structure: examples from three Penstemon (Plantaginaceae) species in the Great Basin. American Journal of Botany 98, 109-121. Kramina, T.E., Sokoloff, D.D., 1999. Taxonomic bearing of stylodium tooth in the genus Lotus (Papilionoideae) with special reference to Lotus creticus L. Feddes Repertorium 110, 527533. Kress, W.J., Wurdack, K.J., Zimmer, E.A., Weigt, L.A., Janzen, D.H., 2005. Use of DNA barcodes to identify flowering plants. Proceedings of the National Academy of Sciences USA 102, 8369-8374. Kunkel, G., 1974. Notes on the genus Heinekenia (Fabaceae) from the Canary Islands. Cuadernos de botanica canaria 22, 7-10. MacArthur, R.H., Diamond, J.M., Karr, J.R., 1972. Density compensation in island faunas. Ecology 53, 330-342. Martín, A.M., Berriel, D.M., 2007. Plan de recuperación del Pico de Paloma (Lotus berthelotii) y el Pico de El Sauzal (Lotus maculatus). Consejeria de Medio Ambiente y Ordenación Territorial, Tenerife, pp. 7065-7993. Masferrer-Arquimbau, R., 1881. Descripcion de la flor y el fruto de Lotus berthelotii (Peliorhynchus berthelotii). Anales de la Sociedad Española de Historia Natural 10, 160.  133  Medina, F.M., 2008. Conservacion del pico de fuego (L. pyranthus) y del picocernicalo (L. eremiticus) en la isla de La Palma. . In: Ambiente, C.d.M. (Ed.). Cabildo Insular de La Palma La Palma. Monod, T., 1980. Contribution a l'étude des Lotus (Papilionaceae) ouest-sahariens et macaronésiens. Adansonia 19, 367-402. O'Leary, T.R., 2009. Phylogenetics and biogeography of Aichryson inferred from chloroplast DNA sequences. University of Arkansas, Kansas, p. 57. Ojeda, L.E., Marrero, G.M., 2004. Lotus berthelotii Masf. In: Bañares, A.B., Blanca, G., Guemes, J., Moreno, J.C., Ortiz, S. (Eds.), Atlas y libro rojo de la flora vascular amenazada de España. Dirección general para la Biodiversidad, Publicaciones del O.A.P.N., Madrid, pp. 390-391. Olesen, J.M., 1985. The Macaronesian bird flower element and its relation to bird and bee opportunistic. Botanical Journal of the Linnean Society 91, 395-414. Olesen, J.M., 1988. Floral biology of the Canarian Echium wildpretii: bird-flower or a water resource to desert bees? Acta Botanica Neerlandica 37, 509-513. Olesen, J.M., Valido, A., 2003a. Bird pollination in Madeira Island. Ardeola 50, 67-69. Olesen, J.M., Valido, A., 2003b. Lizards as pollinators and seed dispersers: an island phenomenon. Trends in Ecology and Evolution 18, 177-181. Oliva-Tejera, F., Caujapé-Castells, J., Naranjo-Suárez, J., Navarro-Déniz, J., Acebes-Ginovés, J.R., Bramwell, D., 2005. Population genetic differentiation in taxa of Lotus (Fabaceae: Loteae) endemic to the Gran Canaria pine forest. Heredity 94, 199-206. Oliva-Tejera, F., Caujapé-Castells, J., Navarro-Déniz, J., Reyes-Betancort, A., Scholz, S., Baccarani-Rosas, M., Cabreara-García, N., 2006. Patterns of genetic divergence on three  134  Canarian endemic Lotus (Fabaceae): implications for the conservation of the endangered L. kunkelii. American Journal of Botany 93, 1116-1124. Ollerton, J., Cranmer, L., Stelzer, R.J., Sullivan, S., Chittka, L., 2009. Bird pollination of Canary Island endemic plants. Naturwissenschaften 96, 221-232. Percy, D.M., Page, R.D.M., Cronk, Q.C.B., 2004. Plant-insect interactions: double-dating associated insect and plant lineages reveals asynchronous radiations. Systematic Biology 53, 120-127. Perez de Paz, P.L., 1990. Lotus pyranthus sp. nov. nuevo endemismo. Vieraea 19, 315-318. Posada, D., 2008. jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution 25, 1253-1256. Rambaut, A., 1996. Se-Al; Sequence Aligment Editor , ver. 1.0 alpha 1. Rodriguez, M.C., Valido, A., 2008. Opportunistic nectar-feeding birds are effective pollinators of bird-flowers from Canary Islands: experimental evidence from Isoplexis canariensis (Scrophulariaceae). American Journal of Botany 95, 1408-1415. Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574. Sandral, G., Remizova, M.V., Sokoloff, D.D., 2006. A taxonomic survey of Lotus section Pedrosia (Leguminosae, Loteae). Wulfenia 13, 97-192. Santos-Guerra, A., 1999. Origen y evolución de la flora canaria. In: Fernández-Palacios, J.M., Bacallado, J.J., Belmonte, J.A. (Eds.), Ecología y Cultura en Canarias. Organismo Autónomo de Museos y Centros, Santa Cruz de Tenerife, pp. 107-129. Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence-based phylogenetic analyses. Systematic Biology 49, 369-381.  135  Sletzer, R., 2005. Sammelstrategien bei Hummeln ein Vergleich zwischen Insel- und Festlandpopulationen Fakultät für Biologie. Universität Würzburg, Würzburg, p. 46. Stiles, F.G., 1978. Ecological and evolutionary implications of bird pollination. American Zoologist 18, 715-727. Swofford, D.L., 2001. PAUP* 4.0: Phylogenetic Analysis Using Parsimony (* and other methods). Sinauer Associates, Sunderland, MA. Valido, A., Dupont, Y.L., Hansen, D.M., 2002. Native birds and insects, and introduced honey bees visiting Echium wildpretii (Boraginaceae) in the Canary Islands. Acta OecologicaInternational Journal of Ecology 23, 413-419. Valido, A., Dupont, Y.L., Olesen, J.M., 2004. Bird-flower interactions in the Macaronesian islands. Journal of Biogeography 31, 1945-1953. Vargas, P., 2007. Are Macaronesian islands refugia of relict plant lineages?: a molecular survey. In: Weiss, S., Ferrand, N. (Eds.), Phylogeography of southern European refugia. Springer, pp. 297-314. Vogel, S., Westerkamp, C., Thiel, B., Gessner, K., 1984. Ornithophilie auf den Canarischen Inseln Plant Systematics and Evolution 146, 225-248. White, J.T., Bruns, T., Lee, S., Taylor, J., 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis, M., Gelfand, D., Snisky, J., White, T.J. (Eds.), PCR protocols: A guide to methods and applications. Academic Press, San Diego, California, pp. 315-322. Zwickl, D.J., 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. The University of Texas at Austin.  136  5 DNA barcoding of plant island radiations and its applicability in species recognition and conservation in Macaronesian Lotus section Pedrosia and Rhyncholotus (Loteae, Leguminosae)1  5.1 Introduction DNA barcoding is a procedure for species identification and recognition that uses a standard sequence of DNA (Hebert et al., 2003). It has proved successful in a variety of animal groups based on the mitochondrial gene cytochrome c oxidase subunit 1 (COI or cox1), which has sufficient variation for species discrimination ( 95%) (Hajibabaei et al., 2006; Kerr et al., 2007; Smith et al., 2008; Ward et al., 2005) in animals, but not plants (Fazekas et al., 2009). Plant barcoding is more complicated as the plant mitochondrial genome evolves at a slower rate than in animals. However, the search for an appropriate barcode region in plants has turned to the plastid genome as this is the fastest evolving genomes and because the plant nuclear genome is difficult to work with in general. A variety of combinations of plastid regions have been tested and such studies have generally found relatively moderate levels (70%) of species discrimination (CBOL, 2009; Fazekas et al., 2008; Fazekas et al., 2009; Kress and Erickson, 2007). After a period of debate and the testing of several regions in a variety of combinations (Kress and Erickson 2007; Sass et al. 2007; Lahaye et al. 2008; Newmaster et al. 2008; Newmaster and Ragupathy 2009; Seberg and Petersen 2009), the suggested barcode regions for plants, matK and rbcL, were recently recommended by the CBOL Plant working group (CBOL, 2009). 1  A version of this chapter will be submitted for publication: Ojeda, I. Santos-Guerra, A., Oliva-  Tejera, F. Jaen Molina, R. Caujape, J. Marrero, A. and Cronk. Q.C.B. BNA barcoding plant island radiations and its applicability in species recognition and conservation in Macaronesian Lotus section Pedrosia and Rhyncholotus (Loteae, Leguminosae). 137  This two-locus combination is able to discriminate about 72% of the samples included in this study to species level with the remaining samples assigned to congeneric species groups. The internal transcribed spacer of the nuclear ribosomal repeat (ITS) has been used as a barcode in fungi (Druzhinina et al. 2005; Kõljalg et al. 2005; Tedersoo et al. 2008). In plants, this region has also been tested in some particular groups (Chase et al. 2005; Kress et al. 2005; Gemeinholzer et al. 2006; Kress and Erickson 2007; Okuyama and Kato 2009). Despite the lower levels of species discrimination obtained in plant barcoding in comparison to animals, it is evident that there are numerous practical applications of plant barcoding in a variety of fields, such as taxonomy, ecology, floristic studies, industry and conservation. However, it is also clear that DNA barcoding in plants will not reach the high levels of species discrimination achieved in animals. Nevertheless, in certain applications is potentially a useful tool for species recognition (Kondo et al., 2007; Kress et al., 2009; Newmaster et al., 2008; Ragupathy et al., 2009; Soininen et al., 2009; Song et al., 2009b). Many of the studies that have tested regions as barcodes in plants have focused their attention on large data sets that span the entirety of land plants or at least angiosperms (CBOL, 2009; Fazekas et al., 2008; Ford et al., 2009; Kress and Erickson, 2007; Lahaye et al., 2008). Their purpose has been the assessment of the universal applicability of the regions in species discrimination. However, it has been argued that the recovery rate of DNA barcode drop in (i) some groups with complex biology, due to a variety of phenomena, such as hybridization, polyploidy, introgression, and (ii) closely related species within the same genus, or in recently evolved groups, such as those of island radiations. To date, the level of species discrimination within the same genus has been tested in a number of cases (Newmaster et al., 2008; Newmaster and Ragupathy, 2009; Sass et al., 2007; Yao et al., 2009). In most studies, the level of species identification within the same genus is promising. However, some of these examples included 138  either few species within the same genus and/or well diverged species groups. There is emerging evidence that at least some closely related groups will be problematic for barcoding (Fazekas et al., 2009; Sass et al., 2007; Seberg and Petersen, 2009; Spooner, 2009). When individual genera have been sampled more extensively, the percentage of species discrimination tends to decrease, even when several regions are combined (Edwards et al., 2008; Kondo et al., 2007; Sass et al., 2007; Seberg and Petersen, 2009; Yao et al., 2009). However, the applicability of species recognition of the recommended barcode regions in very recently evolved groups, such as those of an island radiation, has not being tested extensively. It is unclear if the levels of DNA variation within these two regions, matK + rbcL, will allow species discrimination. Here I present the assessment of five plastid regions suggested as barcodes in previous studies (matK, rpoC1, rpoB, trnH-psbA and rbcL) and the nuclear ribosomal internal transcribed spacer regions, ITS1 and ITS2. Here I tested these regions in a group of Lotus species that radiated in the Macaronesian region (Azores, Madeira, Cape Verde, Canary Islands and Salvage Islands). The group has its highest diversification in the Canary Islands and in mainland Morocco. This Macaronesian Lotus assemblage comprises 41 described species divided into two sections: Pedrosia and Rhyncholotus (Degtjareva et al., 2006). The section Pedrosia comprises 37 recognized species while section Rhyncholotus comprises only four species: Lotus berthelotii, L. maculatus both from Tenerife and L. pyranthus and L. eremiticus from La Palma. It should be noted that section ―Rhyncholotus‖ is in fact nested phylogenetically within section Pedrosia and possibly within a single species (see Chapter 4). The two groups are distinguished from each other by a distinctive flower morphology associated with different pollination syndromes, but within each group vegetative features are more useful for species recognition and identification (Sandral et al., 2006). Species identification  139  using morphology is not always straightforward and a critical examination of a number of morphological features by a specialist will often be necessary for accurate identification. Many species are restricted to specific habitats, such as pine forest and lowland scrub. Furthermore, about 70% of the species are endemic to single islands. Thus, the group is highly susceptible to habitat destruction and at least 10 species are listed under some category of conservation threat, ranging from rare to critically endangered (Bañares et al., 2004; Martín et al., 2008; VV.AA., 2000) (Table 5.1).  5.1.1 Objectives of the study The aims of this study are to answer the following questions: (1) are these DNA barcode regions variable enough to discriminate species in recently evolved groups? (2) is there any potential applicability of these barcoding regions in conservation?, and (3) are these regions variable enough to separate between the two sections Pedrosia and Rhyncholotus?  5.2 Materials and methods 5.2.1 Taxon sampling The sampling includes 78 accessions within the ingroup representing all the species currently described within the sections Pedrosia and Rhyncholotus, except three species (L. loweanus, L. chazalei and L. tibesticus) that were not available for this analysis. In order to sample the intraspecific variation and geographical distribution of some species, I included more than one accession for 27 of the 38 species analyzed (71%). This analysis also included samples from some populations that based on previous molecular and morphological analyses (OlivaTejera et al., 2005; Oliva-Tejera et al., 2006; Sandral et al., 2006), may represent four new  140  undescribed species within the section Pedrosia (Table 5.2). For comparison, I also included five accessions from Lotus section Lotus (Table 5.2).  5.2.2 Selecting barcode regions Several regions have been previously suggested and tested as barcode markers (Fazekas et al., 2008; Kress and Erickson, 2007, 2008a; Kress et al., 2005; Lahaye et al., 2008; Newmaster et al., 2006; Newmaster et al., 2008). For this analysis I included five plastid regions, rbcL, trnHpsbA, matK, rpoB and rpoC1, including the recently recommended two-locus barcode matK + rbcL (CBOL, 2009), along with the nuclear ribosomal ITS region, which has been assessed in some plant groups (Chase et al., 2005; Kress and Erickson, 2007; Kress et al., 2005).  5.2.3. Molecular analysis Genomic DNA was extracted from either fresh leaves, silica-gel dried leaf material or voucher specimens following a modification of the procedure of Doyle and Doyle (1987). Amplification was carried out with the following PCR conditions for all the plastid regions: 94°C for 3 min., 30 cycles of 94°C for 3 min., 45°C for 1 min. and 72°C for 2 min., with a final cycle of 72°C for 5 min. The nuclear ribosomal intergenic spacer ITS was amplified using the following conditions: 94°C for 3 min., 30 cycles of 94°C for 1 min., 55°C for 1 min. and 72°C for 1.5 min, with a final cycle of 72°C for 5 min. Each locus was sequenced and the raw sequence data were imported to Sequencher 4.1 for base-coding, editing and construction of contig sequences. Consensus sequences were imported to Se-Al ver. 1.0 (Rambaut, 1996) and aligning was made manually using conserved regions. Each region was analyzed separately and in two-pair combinations. The analyses were carried both excluding and including missing sequences. In all the gene regions I  141  included the entire sequence, except for a small inversion of 3 bp in the intergenic trnH-psbA region 214-217 position that was excluded from all the analyses.  5.2.4 Assessment of the barcode regions Three parameters have been suggested for the official barcodes: universality, sequence quality and coverage, and discrimination (CBOL, 2009). I evaluated these three parameters in the six regions tested within this group. Amplification success: I estimated the percentage of amplification success on the first trial with the sample set of this group as an indicator of universality, using the same PCR amplification and conditions. Sequence quality and coverage: I estimated the percentage of bidirectional sequences with few or no ambiguous bases. Discrimination: I evaluated discrimination at two levels: species discrimination and discrimination of informal taxonomic groups following according to previous taxonomic analysis based on morphological features (Sandral et al., 2006). I considered that a species was discriminated when it was grouped with samples of the same species (in the case of species with more than one sample) forming a monophyletic group or when it position was resolved fully resolved (in the case of unique samples). Nine informal taxonomic groups at the infrageneric level have been suggested within the Macaronesian assemblage (Table 5.3). I consider that a useful discrimination at this level was achieved when at least 50% of the species were assigned within the same group, forming a monophyletic group.  142  Each region was analyzed separately and in various combinations with Neighbor-joining (NJ) using a Kimura 2-parameter as implemented in PAUP4b10 (Swofford, 2001). This is the standard method used in barcoding studies.  5.3 Results 5.3.1 Universality and sequence quality All regions had above 95% of sequence success and quality, except for the matK region, with 83% of success, which required additional PCR runs to reach this level due to either failure of amplification or due to regions with T or A repeats that caused failure during sequencing. This last barcode region had the lowest level of bidirectional sequence quality (Table 5.5).  5.3.2 Species discrimination of the plastid regions The trnH-psbA region showed the highest level of variation and species discrimination of all regions evaluated (18%), while the rpoB region had the lowest level of variation and species discrimination when analyzed alone (Table 5.4). The combination trnH-psbA + matK showed the highest level of discriminatory power at the species level for two-locus plastid combinations (29%). No differences were observed in species discrimination between the CBOL recommended 2-locus barcode (matK + rbcL) (Fig. 5.1) and other combinations (Table 5.4). In this study, I achieved the identification of 14 species (37%) of the 38 species in this sample when all five plastid regions where combined (Table 5.4). Only two out of 10 species (20%) of conservation concern were identified at the species level (Fig. 5.2). Only five informal taxonomic groups were discriminated with all plastid regions combined (Table 5.4). The intergenic spacer trnH-psbA showed the highest levels of variation and it was the only region where we observed intraspecific variation. I found an inversion (3 bp long) at the 214-217 143  position. The ingroup was either AAA or TTT for this region. This inversion polymorphism is a hairpin loop structure surrounded by an inverted repeat sequence of 22 bp (47 bp in total). Surprisingly 8 out of 27 species with multiple samples showed this polymorphism within a species. This plastid region was also the only one in which we observed indels. Two indels were observed within the ingroup, one large in the position 94-104 that involved 10 bp (TAGATAAAAT) which is shared by all Rhyncholotus species along with four species of Pedrosia; the other indel consisted of a 1 bp deletion (position 194) shared mainly by all members analyzed of the L. assakensis group (L. assakensis, L. creticus and L. pseudocreticus) and most of the members of the L. glaucus group (L. glaucus, L. salvagensis, L. lancerottensis and L. erythrorhyzus) (Table 5.3).  5.3.3 ITS as a barcode in Lotus This region showed the highest level of variability of all regions tested in this study when analyzed alone, with the identification of 26% of the species. The level of species discrimination increased when combined with the plastid regions in pairs. All two-pair combinations of ITS with the plastid regions increased the discriminatory power (Table 5.4). The addition of ITS increased the discriminatory power in this group to 52% of the species when all six regions were combined, including four species with conservation concern (L. arinagensis, L. kunkelii, L. erythrorhyzus and L. genistoides) (Fig. 5.3).  5.4 Discussion Within Macaronesian Lotus I was able to identify 52% of the samples at the species level when all six regions were combined (Table 5.4). Previous studies have reported from 55% (trnHpsbA in Aspalathus) to 92% (e.g. Crocus) of species discrimination in several plant groups 144  (Edwards et al., 2008; Sass et al., 2007; Seberg and Petersen, 2009). As previously suggested, barcodes will have some limitations in closely related species (Chase and Fay, 2009), especially from a rapid and recent island radiation. A recent study in Amazonian trees using a similar amount of barcode regions as in this study found moderate levels of species discrimination (70%) but a tendency to have low levels of species discrimination in species-rich clades (González et al., 2009). The CBOL suggested 2-locus combination (matK + rbcL) discriminated only 18% of the species (Fig. 5.2 and) and no major increase was observed with other two-pair plastid combinations (Table 5.4) My results therefore indicate that the discriminatory power of the barcode regions is low in recently evolved groups, such as those diversifying rapidly as part of an island radiation. Low level of variation has been also reported from Scalesia (Asteraceae) from the Galápagos (Seberg and Petersen, 2009). Similar scenarios with low species discrimination have been reported when barcoding animals on islands, such as the one reported in the genus Copelatus (Coleoptera) in the Fijian archipelago (Monaghan et al., 2006). The applicability of barcodes in the identification of protected or endangered species in this group is low, as I identified four species of the ten (40%) considered under some level of threat. Taken together, these results suggest that DNA barcoding recently evolved groups in islands will remain a challenge for species identification, and are perhaps a ―worst case scenario‖. The performance of the barcodes it is also generally low when applied at the floristic level on islands. Recent evidence from congeneric species in the Garajonay National Park in La Gomera (Jaen-Molina et al. unpubl. data) shows that matK and rbcL in combination only allowed the identification of ca. 75% cases. This result contrasts with other studies which have applied floristic barcoding approaches on diverse areas, such as at La Selva Biological Station (Kress and  145  Erickson, 2008b) and the 50-ha Forest Dynamic Plot on Barro Colorado (Kress et al., 2009), where species identification is above 90%. Despite these results, it is worth noting that the low levels of species identification within the Macaronesian Lotus, the CBOL recommended two-locus barcode will accurately assign samples to section Pedrosia s.l. (Pedrosia + Rhyncholotus) as distinct from the comparison group (Lotus section Lotus). I observed high levels of variation between section Lotus and section Pedrosia s. l. in all barcoding regions. Even the less variable region unequivocally separated the two sections. However, it should be noted that Lotus comprises about 120 species subdivided into 14 sections (Degtjareva et al., 2006), so that further study will be needed to confirm that barcoding can distinguish all sections of Lotus reliably. Another factor that may also decrease species discrimination is the existence of withinspecies variation for some of the analyzed regions. Of the 27 for which I included more than one sample, I observed within-species variation in the intergenic spacer trnH-psbA in eight species (30%) (Table 5.2). This intraspecific variation is similar to that reported for other plant groups in ITS and in some intergenic regions, including the trnH-psbA (Cowan et al., 2006; Edwards et al., 2008; Kress et al., 2009; Seberg and Petersen, 2009; Spooner, 2009; Vischi et al., 2006; Zhang et al., 2009). I only observed intraspecific variation in the intergenic region of the trnH-psbA due to an inversion and deletions, unlike other studies in Aspathalus (Leguminosae) and Helianthus argophylus (Asteraceae) that have reported intraspecific indel variation (Chase and Fay, 2009; Edwards et al., 2008; Vischi et al., 2006).  146  5.4.1 ITS vs. plastid data The low levels of interspecific variation observed in this group in the five plastid regions indicate that additional regions might be considered in particular cases. Faster evolving gene regions, preferably from the nucleus, should therefore be explored in addition to the proposed barcodes. This is supported by the fact that the addition of ITS increased the discriminatory power of these regions in combination with the plastid regions and also when all six regions were combined (Table 5.4 and Fig. 5.3). The discriminatory power for the endangered species also increased with the inclusion of the ITS region. Previous studies have considered ITS as a potential barcode in land plants (Chase et al., 2005; Kress and Erickson, 2007; Kress et al., 2005; Okuyama and Kato, 2009). However, the ITS region has received less attention than plastid regions as a DNA barcode in some groups because of associated problems with sequencing success (González et al., 2009; Kress and Erickson, 2007) and high levels of intraspecific variation in some groups (Edwards et al., 2008; Spooner, 2009). In this analysis I observed intraspecific variation within the ITS region in only 13% of the species (five out of the 38) analyzed. The ITS region performed better than the trnH-psbA in species discrimination despite the intraspecific variation observed, which is probably due to the homogenizing effect of concerted evolution (Álvarez and Wendell, 2003). It is also worth noting that the ITS variability was not homogeneous through the entire ingroup. Whereas it provided sufficient variation for identification in widely distributed species and some (presumably older clades mainly from Africa and Cape Verde) (see Chapter 3), it lacked resolving power in the most recent species groups. However, the general lack of resolution in most phylogenies of Canarian groups undertaken with ITS calls for the exploration of additional regions from the nuclear, especially on groups of recent origin. The ETS and the 5S-NTS regions are less used in phylogenetic studies 147  than the ITS region, but they show higher or similar rates of evolution in some groups where these two regions have been explored (Baldwin and Markos, 1998; Kårehed et al., 2008), and are perhaps good candidates for faster evolving regions within the nuclear genome.  Table 5.1 Macaronesian Lotus species considered under different levels of threat. According to Red List of Spanish Vascular Flora based on the IUCN Red Data Book (IUCN) (VV. AA., 2000), the Atlas of Endangered Spanish Vascular Flora (AESVF) (Bañares et al., 2004), and the ranking according to the Top 100 endangered species of Macaronesia (Martin et al., 2008). Numbers indicate their rank under the Top 100 list, -= not considered within the 100 most endangered species. CR= critically endangered, EN= endangered, VU= vulnerable.  Species  Distribution  L. arinagensis L. berthelotii L. callis-viridis L. dumetorum L. eremiticus L. genistoides L. kunkelii L. maculatus L. mascaensis L. pyranthus L. spartioides  Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands Canary Islands  IUCN 2000  AESVF 2004  Top 100 in Macaronesia  CR CR EN VU CR CR CR VU CR -  CR CR EN CR CR CR CR CR VU  7 25 6 3 -  148  Table 5.2 Species from the sections Pedrosia and Rhyncholotus sampled in this analysis. Distribution: G= La Gomera, P=La Palma, T=Tenerife, GC= Gran Canaria, CV= Cape Verde, M= Madeira, H= Hierro, L= Lanzarote, F= Fuerteventura. UBC= University of British Columbia, LBCVC= Jardín Botánico Canario Viera y Clavijo, JAO= Jardín de Aclimatación de la Orotava.  149  Taxon Outgroup  Collection information  Voucher, herbarium  Lotus japonicus MG20 (Regel) K. Larsen  Cultivated from seeds at UBC  Ojeda 69/UBC  Lotus japonicus Gifu B-129 (Regel) K. Larsen Lotus filicaulis Durieu Lotus corniculatus L. Lotus burttii Borsos  Cultivated from seeds at UBC  Ojeda 70/UBC  Cultivated from seeds at UBC Vancouver, BC Cultivated from seeds at UBC from Univ. Miyazaki  Ojeda 71/UBC Ojeda 46/UBC Ojeda 72/UBC  Cultivated JBCVC # 164/06 Sao Nicolao CV Cultivated UBC # PI 631779, Casablanca, Morocco Cultivated UBC # PI 631956, Kourigba, Morocco DNA bank # 651 Barranco Viejo, Arinaga, GC DNA bank # 652 Ctra faro, Arinaga, GC Cult. JCVC # 5435/UDH/07 Punta de Pargo, M Voucher, Punta San Lorenzo, M Voucher, Tarfaya-Tan Tan Sahara, Africa DNA Bank # 1084 Tiznir, Aglou Plage, Morocco  Ojeda 180/ UBC Ojeda 78/UBC José Cruz & Alicia Roca Felicia Oliva & José Naranjo Ojeda 189/UBC ORT # 37806 Molero 1992 (Fernández Casas 13699) ORT # 36336 Ojeda 182/UBC Ojeda 181/UBC Alicia Roca & Bernardo Navarro Ojeda 169/UBC Ojeda 206/UBC Ojeda 210/UBC Ojeda 188/UBC Ojeda 242/UBC Ojeda 213/UBC Ojeda 228/UBC Ojeda 207/UBC Ojeda 209/UBC Ojeda 244/UBC Ojeda 243/UBC A. Santos Felicia Oliva & José Naranjo/JBCVC Ojeda 187/UBC Ojeda 233/UBC José Naranjo & Paloma MayaOjeda 198/UBC F. Oliva, J. Naranjo, J. Navarro, I. Santana & B. Vilches/JBCVC Marrero et al/JBCVC Marrero et al/JBCVC S.L. Jury & T.M. Upson 20503/RNG S.L. Jury & T.M. Upson 20480/RNG José Cruz & Miguel Alemán Ojeda 176/UBC ORT # 36458 Marrero et al/JBCVC Ojeda 170/UBC A. Santos ORT # 33596 ORT # 36675 S.L. Jury 14471/RNG Fernández Casas 13737/RNG José Cruz & Ruth Jaén/JBCVC Ojeda 200/UBC F. Damblon 84/40/RNG Davies 53484/RNG ORT # 36670 ORT # 35118 ORT # 35116  Ingroup Lotus section Pedrosia (Lowe) Christ. Lotus arborescens Lowe ex Cout. Lotus arenarius Brot. Lotus arinagensis Brawm. Lotus argyrodes R.P Murray Lotus assakensis Brand  Lotus azoricus P. W. Ball Lotus bollei Lotus brunneri Webb in Hooker Lotus callis-viridis Bramwell & D.H. Davis  Cultivated JAO # 161-00, Azores Cultivated JBCVC # 163/06 Sao Vicente, CV Cultivated JBCVC # 514B/07 Sal, CV DNA Bank # 654 Andén Verde, GC Andén Verde, GC  Lotus campylocladus Webb & Berthel.  Road to Cañada Teide, T Arona-Ifonche, T Cultivated JBCVC # 64/05 Cultivated UBC, PI 505409, Spain Mirador Jardina, Mercedes, Anaga, T Teno Alto, Teno, T Inchereda, G Epina, G Cultivated UBC # PI 631959, Ouarzate, Morocco Cultivated UBC # PI 631784, Tiznir, Morocco Fuerteventura DNA Bank # 655 Cañadón Sombrío, GC  Lotus creticus L. Lotus dumetorum Webb ex R. P. Murray Lotus emeroides R. P. Murray Lotus eriosolen (Maire) Mader & Podlech L. erythrorhyzus Bolle L. genistoides (nom. nudum) Lotus glaucus Sol. Lotus hillebrandii Christ Lotus holosericeus Webb & Berthel. Lotus jacobaeus L. Lotus jolyi Battand.  Lotus kunkelii (Esteve) Bramwell & D. H. Davis Lotus lancerottensis Webb & Berth. Lotus latifolius Brand Lotus leptophyllus (Lowe) K. Larsen Lotus macranthus Lowe Lotus maroccanus Ball Lotus mascaensis Burchard Lotus pseudocreticus Maire, Weiller & Wilczek Lotus purpureus Webb Lotus salvagensis R.P. Murray  Cult. JBCVC # 223/B/07 Porto Nurbita (??), M Cult. JAO 19-05 DNA Bank # 656 Llanos Chozas, P Mirador Isora, H DNA Bank # 657 Pilancones, GC DNA Bank # 658 Bordeira bei Piorno Fogo CV DNA Bank # 2089 Ribeira Monte espia Fogo CV Voucher, Province Tan Tan, Morocco Voucher, Province Guelmin, Morocco  DNA Bank # 3805, Barranco Jinamar, GC Cult. JBCVC # 217/07 DNA Bank # 3823 Villaverde, Betancuira, F Voucher, L DNA bank # 1812 Crtra. Porto Novo, CV Barranco Guayedra, GC GC Voucher, Pico Branco, Porto Santo, M Voucher, M Voucher, Talouine, Morocco Voucher, Marrakech, Morocco DNA Bank # 659, Cultivated JBCVC Valle de Masca, T Voucher, SW Agadir Morocco Voucher, Tamri Agadir, Morocco Cult. JAO # 130-99 Voucher, Salvage Grande, SGVoucher, Acantilados del NE, Salvage Islands  150  Taxon  Collection information  Voucher, herbarium  Lotus sessilifolius D.C. subsp. villossisimus (Pitard) Sandral & Sokoloff  Las Playas, S from Parador, H  Ojeda 196/UBC  Lotus sessilifolius D.C. subsp. sessilifolius  Poris de Abona, T Puntallana, G Playa Pocito, Mazo, P San Juan-Guía de Isora, T  Ojeda 225/UBC Ojeda 208/UBC A. Santos Ojeda 205/UBC  Tamadaba, GC Pinar Pajonales, GC DNA bank # 662 Chira-Pinar Santiago, GC Presa las Niñas, GC Garachico, Ermita San Roque, T  Ojeda 217/UBC Ojeda 216/UBC F. Oliva, J. Naranjo, J. Navarro, I. Santana & B. Vilches/JBCVC Ojeda 211/UBC Ojeda 446/UBC  Punta Hidalgo, T South Roque Dos hermanos, Anaga, T Teno Alto, T Punta Góngora, GC Cortijo de San Ignacio, GC Punta Teno, Teno, T  Ojeda 193/UBC Ojeda 194/UBC Ojeda 167/UBC Ojeda 203/UBC Ojeda 230/UBC  Ifonche, T Cultivated UBC, commercial plant DNA Bank # 3839, cult. JBCVC (366/04), Garafia, P DNA Bank # 660, ex horto (8/04) Puertito Sauzal, T Cult. UBC, commercial plant DNA Bank # 661 JBCVC 210/99 DNA Bank # 3842, Cult. Vivero Ceplam  Ojeda 238/UBC Jose Cruz  Lotus sessilifolius DC. var. pentaphyllus (Link) D. H. Davis Lotus spartioides Webb & Berthel.  Lotus tenellus (R. Lowe) Sandral, Santos & D.D. Sokoloff Lotus sp. nov. ined. 1  Lotus sp. nov. ined. 2 (L. leptophyllus group) Lotus sp. nov. ined. 3 (L. spartioides group) Lotus sp. nov. ined. 4 (L. sessilifolius group) Lotus section Rhyncholotus (Monod) D.D. Sokoloff L. berthelotii Masf. var. berthelotii Lotus eremiticus A. Santos L. maculatus Breitf. Lotus pyranthus P. Perez  Rafael Almeida/JBCVC Ojeda 239 Ojeda 175/UBC -  151  Table 5.3 Sections Pedrosia and Rhyncholotus and their informal classification based on morphological features. *According to Sandral et al. (2006). ¶Species not sampled in this analysis.  Section  Informal taxonomic groups below section level Lotus purpureus group  Lotus arenarius group*  Lotus sessilifolius group*  Pedrosia Lotus argyrodes group*  Lotus campylocladus group*  Lotus glaucus group*  Lotus assakensis group*  Lotus jolyi group* Rhyncholotus  Rhyncholotus group*  Species L. arborescens L. bollei L. brunneri L. jacobaeus L. purpureus L. latifolius L. arenarius L. maroccanus L. eriosolen L. sessilifolius L. mascäensis L. arinagensis L. emeroides L. kunkelii L. argyrodes L. macranthus L. azoricus L. loweanus¶ L. callis-viridis L. campylocladus L. genistoides L. holosericeus L. hillebrandii L. spartioides L. dumetorum L. glaucus L. tenellus L. leptophyllus L. salvagensis L. lancerottensis L. erythrorhyzus L. assakensis L. creticus L. pseudocreticus L. chazalei¶ L. jolyi L. tibesticus¶ L. berthelotii L. eremiticus L. maculatus L. pyranthus  Distribution Cape Verde  Africa, Spain  Canary Islands  Azores, Madeira  Canary Islands  Canary Islands, Salvage Islands and Madeira  Africa, Mediterranean  Africa Canary Islands  152  Table 5.4 Performance of the five plastid regions and the nuclear ribosomal ITS tested separate and in two-pair combinations. A= including all accessions, B= excluding accessions with missing sequences in two-pair combinations. * Informal sections according to Sandral et al. (2006)  One region ITS trnH-psbA matK rpoC1 rbcL rpoB Plastid combinations matK + trnH-psbA matK + rpoC1 rpoC1 + trnH-psbA rbcL + trnH-psbA matK + rbcL matK + rpoB rpoB + trnH-psbA rbcL + rpoC1 rpoB + rpoC1 rbcL + rpoB All plastid combined ITS + plastid ITS + trnH-psbA ITS + rpoC1 ITS + matK ITS + rpoB ITS + rbcL All six regions combined  Aligned sequence (bp)  No. of species discriminated/ endangered  No. of informal taxonomic groups*  621 342 867 511 588 354  10/0 7/1 7/1 5/1 2/0 0/0  3 4 4 0 0 0  1209 1378 853 930 1455 1221 696 1099 865 942 2662  A 11/2 10/2 10/1 7/1 7/1 6/1 5/0 3/0 5/1 4/0 9/2  B 13/2 10/3 9/0 9/1 7/1 6/0 6/0 3/0 5/1 3/0 14/3  A 4 4 3 3 3 3 4 0 0 1 4  B 4 3 3 3 2 2 4 0 0 0 4  963 1132 1468 975 1209 3283  15/3 12/1 11/1 11/1 10/0 20/4  14/3 11/1 7/0 11/1 9/0 17/3  4 3 4 3 3 4  4 3 4 3 3 4  153  Table 5.5 Nuclear and plastid gene regions tested in this analysis with their specific primers and performance. Region trnH-psbA matK rpoC1 rbcL rpoB ITS  Primer pair Fw PA Rev TH matK2.1F matK3.2X rpoC1F rpoC14R 80F ajf634R1 rpoB2F rpoB3R ITS4 ITS5  PCR success 96  Sequencing success 98  No. indels 2  83  85  0  96  100  0  97  100  0  97  100  0  100  99  2  154  Figure 5.1 NJ tree generated with the combination of the CBOL recommended two-locus, matK + rbcL. Gray squares represent species with more than one sample and species in a square represent species with a single accession. Branches with a black square represent informal taxonomic groups identified. Species in bold belong to section Rhyncholotus while species not in bold are included within section Pedrosia. *Endangered species identified.  155  L. japonicus GIFU B129 L. japonicus MG20 L. corniculatus  0.0005 substitutions/site  L. arborescens L. brunneri L. jacobaeus L. jacobaeus L. arenarius L. maroccanus L. arenarius L. arinagensis Gran Canaria * L. arinagensis Gran Canaria L. assakensis L. creticus L. assakensis L. erythrorhyzus L. kunkelii L. glaucus L. glaucus L. kunkelii L. lancerottensis L. lancerottensis L. argyrodes L. argyrodes L. azoricus L. salvagensis L. bollei L. latifolius L. jolyi L. jolyi L. eriosolen L. eriosolen L. spartioides Presa Las Niñas L. callis-viridis Andén Verde, Gran Canaria L. campylocladus Arona-Ifonche, Tenerife L. emeroides Epina, La Gomera L. hillebrandii Llanos Chozas, La Palma L. holosericeus Gran Canaria L. leptophyllus Barranco Guayedra, Gran Canaria L. mascaensis Punta Teno, Tenerife Lotus sp. nov. 1 Punta Hidalgo, Tenerife L. berthelotii Tafira, La Florida, Tenerife L. maculatus L. mascaensis Valle Masca, Tenerife L. sessilifolius subsp. sessilifolius Punta Llana, Gomera L. sessilifolius subsp. sessilifolius Poris de Abona Lotus sp. nov. 1 Roque dos Hermanos, Tenerife L. berthelotii L. spartioides Chira-Pinar Santiago L berthelotii Ifonche L. maculatus L. pyranthus L. pyranthus L. eremiticus Lotus sp. nov. 1 Teno Alto, Tenerife L sessilifolius var. pentaphyllus L. maroccanus L. leptophyllus L hillebrandii L. genistoides L. emeroides L. campylocladus L. callis-viridis L. mascaensis Valle de Masca, Tenerife  L. burtii L. filicaulis  156  Figure 5.2 NJ tree generated with the combination of all five plastid regions. Gray squares represent species with more than one sample and species in a square with a single accession. Branches with a black square represent informal taxonomic groups identified. Species in bold belong to section Rhyncholotus while species not in bold are included within section Pedrosia. *Endangered species identified.  157  L. japonicus GIFU B129 L. japonicus MG20 L. corniculatus  0.0001 substitutions/site  L. arenarius L. maroccanus L. arenarius L. maroccanus L. brunneri L. jacobaeus L. jacobaeus L. eriosolen L. eriosolen L. jolyi L. jolyi L. latifolius L. salvagensis L. arinagensis * L. arinagensis L. assakensis L. assakensis L. erythrorhyzus L. lancerottensis L. lancerottensis L. kunkelii L. glaucus L. glaucus L. creticus L. kunkelii L. campylocladus Lotus sp. nov. 1 R. dos Hermanos, Tenerife L. campylocladus L. hillebrandii L. callis-viridis L. callis-viridis L. emeroides L. emeroides L. genistoides L. hillebrandii L. holosericeus L. mascaensis Valle Masca, Tenerife L. spartioides Chira-Pinar Santiago L. sessilifolius var. pentaphyllus Tenerife L. berthelotii L. sessilifolius subsp. sessilifolius La Gomera L. pyranthus L. pyranthus L. leptophyllus Gran Canaria L. leptophyllus Gran Canaria L. sessilifolius subsp. sessilifolius La Palma L. mascaensis Punta Teno, Tenerife Lotus sp. nov. 1 Teno Alto, Tenerife L. berthelotii L. maculatus L. maculatus L. eremiticus L. berthelotii Ifonche, Tenerife Lotus sp. nov. 1 Punta Hidalgo, Tenerife L. argyrodes L. argyrodes L. azoricus L. spartioides Tamadaba * L. spartioides Presa Las Niñas L. spartioides Pinar Pajonales L. mascaensis  L burtii L filicaulis  158  Figure 5.3 NJ tree generated with the combination of all six regions tested (rbcL, matK, trnH-psbA, rpoC1, rpoB and ITS). Gray squares represent species with more than one sample and species in a square represent species with a single accession. Branches with a black square represent informal taxonomic groups identified. Species in bold belong to section Rhyncholotus while species not in bold are included within section Pedrosia. *Endangered species identified.  159  L. japonicus GIFU B129 L. japonicus MG20 L. burtii L. filicaulis L. corniculatus  0.0005 substitutions/site  L. arenarius L. arenarius L. brunneri L. hillebrandii L. holosericeus L. lancerottensis L. emeroides L. emeroides L. arinagensis * L. arinagensis L. jolyi L. jolyi L. salvagensis L. maroccanus L. maroccanus L. mascaensis L. sessilifolius subsp. sessilifolius L sessilifolius subsp. sessilifolius L. eremiticus L. pyranthus L. berthelotii L. berthelotii L. pyranthus L. berthelotii L. maculatus L. maculatus L. argyrodes L. argyrodes L. azoricus L. spartioides Presa Niñas L. leptophyllus L. leptophyllus Lotus sp. nov. 1 Teno, Tenerife Lotus sp. nov. 1 Anaga, Tenerife L. assakensis L. assakensis L. erythrorhyzus * L. lancerottensis L. glaucus L. glaucus L. kunkelii * L. kunkelii L creticus L. callis-viridis L. sessilifolius var. pentaphyllus L. callis-viridis L. hillebrandii L. eriosolen L. eriosolen L. latifolius L genistoides * L. spartioides Tamadaba L. mascaensis Lotus sp. nov. 1 Roque dos Hermanos L. campylocladus L. campylocladus L. mascaensis L. jacobaeus L. jacobaeus  160  5.5 Bibliography Álvarez, I., Wendell, J.F., 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molecular Phylogenetics and Evolution 29, 417-434. Baldwin, G., Markos, S., 1998. Phylogenetic utility of the external transcribed spacer (ETS) of 18S–26S rDNA: congruence of ETS and ITS trees of Calycadenia (Compositae). Molecular Phylogenetics and Evolution 10, 449–463. Bañares, A., Blanca, G., Güemes, J., Moreno, J.C., Ortiz, S., 2004. Atlas y Libro Rojo De La Flora Vascular Amenazada de España. Dirección General de Conservación de la Naturaleza, Madrid. CBOL, 2009. A DNA barcode for land plants. Proceedings of the National Academy of Sciences of the USA 106, 12794-12797. Chase, M.W., Fay, M.F., 2009. Ecology. Barcoding of plants and fungi. Science 325, 682-683. Chase, M.W., Salamin, N., Wilkinson, M., Dunwell, J.M., Kesanakurthi, R.P., Haidar, N., Savolainen, V., 2005. Land plants and DNA barcodes: short-term and long-term goals. Philosophical Transactions of the Royal Society B: Biological Sciences 360, 1889-1895. Cowan, R.S., Chase, M.W., Kress, W.J., Savolainen, V., 2006. 300, 000 species to identify: problems, progress and prospects in DNA barcoding of land plants. Taxon 55, 611-616. Degtjareva, G.V., Kramina, T.E., Sokoloff, D.D., Samigullin, T.H., Valiejo-Roman, C.M., Antonov, A.S., 2006. Phylogeny of the genus Lotus (Leguminosae, Loteae): evidence from nrITS sequences and morphology. Canadian Journal of Botany 84, 813-830. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19, 11-15. Edwards, D., Horn, A., Taylor, D., Savolainen, V., Hawkins, J.A., 2008. DNA barcoding of a large genus, Aspalathus L. (Fabaceae). Taxon 57, 1317-1327.  161  Fazekas, A.J., Burgess, K.S., Kesanakurti, P.R., Graham, S.W., Newmaster, S.G., Husband, B.C., Percy, D.M., Hajibabaei, M., Barrett, S.C., 2008. Multiple multilocus DNA barcodes from the plastid genome discriminate plant species equally well. PLoS ONE 3, e2802. Fazekas, A.J., Kesanakurti, P.R., Burgess, K.S., Percy, D.M., Graham, S.W., Barrett, S.C., Newmaster, S.G., Hajibabaei, M., Husband, B.C., 2009. Are plant species inherently harder to discriminate than animal species using DNA barcoding markers? Molecular Ecology Resources 9, 130-139. Ford, C.S., Ayres, K.L., Toomey, N., Haider, N., AIDER4, Stahl, J.V.A., Kelly, L.J., Wikstrom, N., Hollingsworth, P.M., Duff, J., Hoot, S.B., OOT9, Cowan, R.S., Chase, M.W., Wilkinson, M.J., 2009. Selection of candidate coding DNA barcoding regions for use on land plants. Botanical Journal of the Linnean Society 159, 1-11. González, M.A., Baraloto, C., Engel, J., Mori, S.A., Petronelli, P., 2009. Identification of Amazonian Trees with DNA Barcodes. PLoS ONE 4, e7483. Hajibabaei, M., Janzen, D.H., Burns, J.M., Hallwachs, W., Hebert, P.D., 2006. DNA barcodes distinguish species of tropical Lepidoptera. Proceedings of the National Academy of Sciences, USA 103, 968-971. Hebert, P.D., Cywinska, A., Ball, S.L., deWaard, J.R., 2003. Biological identifications through DNA barcodes. Proceedings of the Royal Society of London. Series B, Biological Sciences 270, 313-321. Kårehed, J., Groeninckx, I., Dessein, S., Motley, T.J., Bremer, B., 2008. The phylogenetic utility of chloroplast and nuclear DNA markers and the phylogeny of the Rubiaceae tribe Spermacoceae. Molecular Phylogenetics and Evolution 49, 843-866.  162  Kerr, K.C., Stoeckle, M.Y., Dove, C.J., Weigt, L.A., Francis, C.M., Hebert, P.D., 2007. Comprehensive DNA barcode coverage of North American birds. Molecular Ecology Notes 7, 535-543. Kondo, K., Shiba, M., Yamaji, H., Morota, T., Zhengmin, C., Huixia, P., Shoyama, Y., 2007. Species identification of licorice using nrDNA and cpDNA genetic markers. Biological & Pharmaceutical Bulletin 30, 1497-1502. Kress, W.J., Erickson, D.L., 2007. A two-locus global DNA barcode for land plants: the coding rbcL gene complements the non-coding trnH-psbA spacer region. PLoS ONE 2, e508. Kress, W.J., Erickson, D.L., 2008a. DNA barcodes: genes, genomics, and bioinformatics. Proceedings of the National Academy of Sciences USA 105, 2761-2762. Kress, W.J., Erickson, D.L., 2008b. DNA Barcoding—a Windfall for Tropical Biology? Biotropica 40, 405-408. Kress, W.J., Erickson, D.L., Jones, F.A., Swensond, N.J., Perez, P., Sanjurb, O., Berminghamb, E., 2009. Plant DNA barcodes and a community phylogeny of a tropical forest dynamics plot in Panama. Proceedings of the National Academy of Sciences USA 106, 1862118626. Kress, W.J., Wurdack, K.J., Zimmer, E.A., Weigt, L.A., Janzen, D.H., 2005. Use of DNA barcodes to identify flowering plants. Proceedings of the National Academy of Sciences USA 102, 8369-8374. Lahaye, R., van der Bank, M., Bogarin, D., Warner, J., Pupulin, F., Gigot, G., Maurin, O., Duthoit, S., Barraclough, T.G., Savolainen, V., 2008. DNA barcoding the floras of biodiversity hotspots. Proceedings of the National Academy of Sciences USA 105, 29232928.  163  Martín, J.L., Arechavaleta, M., Borges, P.A., Faria, B., 2008. Top 100. Las 100 especies amenazadas prioritarias de gestión en la región europea biogeográfica de la Macaronesia. . Consejería de Medio Ambiente y Ordenación Territorial, Gobierno de Canarias., Gran Canaria. Monaghan, M.T., Balke, M., Pons, J., Vogler, A.P., 2006. Beyond barcodes: complex DNA taxonomy of a South Pacific Island radiation. Proceedings of the Royal Society of London. Series B, Biological Sciences 273, 887-893. Newmaster, S.G., Fazekas, A.J., Ragupathy, S., 2006. DNA barcoding in the land plants: evaluation of the rbcL in a multigene tiered approach. Canadian Journal of Botany 84, 335-341. Newmaster, S.G., Fazekas, A.J., Steeves, R., Janovec, J., 2008. Testing candidate plant barcode regions in the Myristicaceae. Molecular Ecology Resources 8, 480-490. Newmaster, S.G., Ragupathy, S., 2009. Testing plant barcoding in a sister species complex of pantropical Acacia (Mimosideae, Fabaceae). Molecular Ecology Resources 9, 172-180. Okuyama, Y., Kato, M., 2009. Unveiling cryptic species diversity of flowering plants: successful biological species identification of Asian Mitella using nuclear ribosomal DNA sequences. BMC Evolutionary Biology 9, 105. Oliva-Tejera, F., Caujapé-Castells, J., Naranjo-Suárez, J., Navarro-Déniz, J., Acebes-Ginovés, J.R., Bramwell, D., 2005. Population genetic differentiation in taxa of Lotus (Fabaceae: Loteae) endemic to the Gran Canaria pine forest. Heredity 94, 199-206. Oliva-Tejera, F., Caujapé-Castells, J., Navarro-Déniz, J., Reyes-Betancort, A., Scholz, S., Baccarani-Rosas, M., Cabreara-García, N., 2006. Patterns of genetic divergence on three Canarian endemic Lotus (Fabaceae): implications for the conservation of the endangered L. kunkelii. American Journal of Botany 93, 1116-1124. 164  Ragupathy, S., Newmaster, S.G., Murugesan, M., Balasubramaniam, V., 2009. DNA barcoding discriminates a new cryptic grass species revealed in an ethnobotany study by the hill tribes of the Western Ghats in southern India. Molecular Ecology Resources 9 (Suppl. 1), 164-171. Rambaut, A., 1996. Se-Al; Sequence Aligment Editor , ver. 1.0 alpha 1. Sandral, G., Remizova, M.V., Sokoloff, D.D., 2006. A taxonomic survey of Lotus section Pedrosia (Leguminosae, Loteae). Wulfenia 13, 97-192. Sass, C., Little, D.P., Stevenson, D.W., Specht, C.D., 2007. DNA barcoding in the cycadales: testing the potential of proposed barcoding markers for species identification of cycads. PLoS ONE 2, e1154. Seberg, O., Petersen, G., 2009. How many loci does it take to DNA barcode a crocus? PLoS ONE 4, e4598. Smith, M.A., Poyarkov, N.A., Hebert, P.D., 2008. CO1 DNA barcoding amphibians: take the chance, meet the challenge. Molecular Ecology Resources 8, 235-246. Soininen, E.M., Valentini, A., Coissac, E., Miquel, C., Gielly, L., Brochmann, C., Brysting, A.K., Sønstebø, J.H., Ims, R.A., Yoccoz, N.G., Taberlet, P., 2009. Analyzing diet of small herbivores: the efficiency of DNA barcoding coupled with high-throughput pyrosequencing for deciphering the composition of complex plant mixtures. Frontiers in Zoology 6, 1-9. Song, J., Yao, H., Li, Y., Li, X., Lin, Y., Liu, C., Han, J., Xie, C., Chen, S., 2009. Authentication of the family Polygonaceae in Chinese pharmacopoeia by DNA barcoding technique. Journal of Ethnopharmacology 124, 434-439. Spooner, D.M., 2009. DNA barcoding will frequently fail in complicated groups: An example in wild potatoes. American Journal of Botany 96, 1176-1189. 165  Swofford, D.L., 2001. PAUP* 4.0: Phylogenetic Analysis Using Parsimony (* and other methods). Sinauer Associates, Sunderland, MA. Vischi, M., Arzenton, F., De Paoli, E., Paselli, S., Tomat, E., Olivieri, A.M., 2006. Identification of wild species of sunflower by a specific plastid DNA sequence. Helia 29, 11-18. VV.AA., 2000. Lista Roja de Flora Vascular Española (valoracion segun categorias UICN). Conservación Vegetal 6 (Extra), 11-38. Ward, R.D., Zemlak, T.S., Innes, B.H., Last, P.R., Hebert, P.D., 2005. DNA barcoding Australia's fish species. Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences 360, 1847-1857. Yao, H., Song, J.Y., Ma, X.Y., Liu, C., Li, Y., Xu, H.X., Han, J.P., Duan, L.S., Chen, S.L., 2009. Identification of Dendrobium species by a candidate DNA barcode sequence: The chloroplast psbA-trnH intergenic region. Planta Medica 75, 667-669. Zhang, J., Wang, J., Xia, T., Zhou, S., 2009. DNA barcoding: species delimitation in tree peonies. Science in China Series C: Life Sciences 52, 568-578.  166  6 Temporal, but not spatial, changes in expression patterns of floral identity genes are associated with the evolution of bird pollination in Lotus (Leguminosae)1  6.1 Introduction The understanding of the genetic and molecular basis of floral symmetry and petal identity has increased in recent years using model species (Feng et al., 2006b; Luo et al., 1996; Wang et al., 2008) and this knowledge is now being extended to non-model species. In particular CYCLOIDEA, a transcription factor that belongs to the Class II TCP family, is known to have a major role in petal identity in the Leguminosae. In legumes this gene family has four members at least two of these genes play a role in the determination of petal identity (Citerne et al., 2003; Feng et al., 2006b; Wang et al., 2008). Lotus japonicus CYCLOIDEA2 (LjCYC2) specifies dorsal petal identity and LjCYC3 acts to specify the identity of lateral petals. Each one of these genes is associated with a particular epidermal micromorphology, the former with papillose conical cells (PCS) and the latter with tabular rugose cells (TRS) (Chapter 4) The tribe Loteae has been identified as one of the groups with a highly differentiated distribution of epidermal types, in which a particular epidermal type is highly associated with a specific petal identity (Ojeda et al., 2009).  1  A version of this chapter will be submitted for publication: Ojeda, I. Santos-Guerra, A., Oliva-  Tejera, F. Jaen Molina, R. Caujape, J. Marrero, A. and Cronk. Q. C. B. Temporal, but not spatial, changes in expression patterns of floral identity genes are associated with the evolution of bird pollination in Lotus (Leguminosae).  167  Homologues of CYCLOIDEA have formerly been associated with shifts in floral symmetry and pollination syndrome (Citerne et al., 2006). Shifts of pollinators require the modification of several floral features associated with the attraction and reward of legitimate pollinators, or exclusion of illegitimate pollinators. Evolutionary change in pollination syndrome involves morphological aspects of the flower (such as flower colour, flower size, length of nectar spurs, scent production, and nectar volume and composition). In a few cases, the specific genes involved in the transitions have been identified, such as Anthocyanin 2 (AN2) and flavonoid-3’-hydroxylase (F3’H), or major quantitative trait loci (QTL) (Bradshaw et al., 1998; Bradshaw and Schemske, 2003; Cronk and Ojeda, 2008; Galliot et al., 2006a; Hoballah et al., 2007; Quattrochio et al., 1999; Stuurman et al., 2004; Whittall and Hodges, 2007; Zufall and Rausher, 2004). Here I study a transition in flower morphology from melittophilous ancestors to ornithophily (specifically pollination by opportunistic passerine birds) that evolved in a group of four species within the genus Lotus section Rhyncholotus (Fig. 6.1A-D) in the islands of Tenerife and La Palma in the Canary Island archipelago. There are no modifications in flower symmetry during the evolutionary transition between these two pollination syndromes (both groups have zygomorphic flowers). The major differences between the four ornithophilous compared with the melittophilous species are the size and shape of the petals and their orientation within the flower (Fig. 6.1E-H, and Fig. 6.2D), differences in flower colour and petal micromorphology. These modifications suggest changes in relative growth rates and in the role of individual petal types on pollination attraction. The pollination mechanism in the melittophilous species is identical to that described for L. corniculatus (Proctor et al., 1996). Pollination takes place when the insect lands and depresses the wings and the keel, forcing out a string of pollen from the stamens located within the keel and 168  placing it in the underside of the visitor (Proctor et al., 1996). Bees need a landing platform and the horizontal orientation of the flowers allows the lateral petals to play this role (Fig. 6.2A, F). On the other hand, in the ornithophilous species the pollinator seeks nectar located at the base of the calyx and the pollen is deposited either on the top of the head or in the throat when the dorsal petal is pressed down (Fig. 6.2C, E) (Olesen, 1985). In this case, the bird needs no landing platform (as it usually forages from the ground) and a flower in vertical orientation is better suited for nectar storage and pollen placement (Fig. 6.2E). To date, only the Canarian chiffchaff (Phylloscopus canariensis) and the blue tit (Parus caeruleus), have been identified foraging in two of these four species (Ollerton et al., 2008; Stelzer, 2005). The petal modifications in these four bird-pollinated species seem remarkable within the tribe Loteae, a group that comprises about 275 species that predominantly have a bee-pollination syndrome and a fairly uniform flower morphology. To my knowledge, this is the only confirmed case of bird pollination within this tribe. There is another report of a Costa hummingbird visiting Hosackia rigida, but the flower morphology of this species is not ornithophilous and these birds visit this species only in the absence of the other bird-pollinated flowers (Grant and Grant, 1968). Phylogenetic analyses in this group (Allan et al., 2004; Sandral et al., 2006) and geological evidence from La Palma and Tenerife (Carracedo et al., 2002) suggest that this pollination syndrome probably evolved recently (see Chapter 4).  5.1.1 Objectives of the study The objectives of this chapter are: 1) to determine the distribution of epidermal types on petals of the sections Pedrosia and the rhyncholotus group and its relation to modification in pollinator syndromes, and 2) to analyze the association of changes in the expression of petal identity genes with alterations in petal morphology that are related to shifts in pollination. 169  6.2 Material and methods 6.2.1 Plant material and growth conditions The following five species were used in the gene expression analysis: Lotus japonicus ecotype Gifu B129, L. filicaulis (bee-pollinated), and L. sessilifolius (bee-pollinated but closely related to the bird-pollinated species), L. maculatus and L. berthelotii (bird-pollinated). The plants were grown in the nursery in pots of 10-20 cm in diam. at 20-25 °C and were more than 6 weeks old when flowers were collected for analyses. Lotus berthelotii and L. maculatus were purchased from commercial nurseries. They were grown under the same environmental conditions as the rest of the species, except during the vernalization period of 30 days in which the temperature was reduced (1 -10°C). After this period, L. berthelotii and L. maculatus started to bloom from MayAugust. Lotus sessilifolius was propagated from seeds collected from the field in Tenerife. For the petal micromorphology survey I analyzed a total of seven genera and 56 species within the tribe Loteae. Within Lotus, I analyzed representative species of 9 (out of 14) sections currently recognized within this group (Degtjareva et al., 2006). This analysis included all four species with a bird pollination syndrome and 32 (out of 36) of the currently recognized species within section Pedrosia with a bee pollination syndrome (Table 6.1) (Sandral et al., 2006).  6.2.2 Flower developmental stages and petal growth Early flower development in L. japonicus has previously been divided into seven stages from stage 0 when floral primordia are initiated to stage 7 when the primordia of all five petals are initiated but elongation has not started (Dong et al., 2005). At this stage the veins and the characteristic epidermal type of each petal type has not differentiated (Dong et al., 2005; Feng et al., 2006b; Zhang et al., 2003). For this study I extended the classification of L. japonicus flower development into a further six stages, 8-13 (Fig. 6.1). I used morphological landmarks and size 170  relationships of the petals to characterize each stage. Further, analogous stages were established on the Pedrosia and the rhyncholotus group analyzed following the same landmarks. At stage 8 no petal is visible as they are covered by the sepals. Differentiation of veins and of the characteristic epidermal type of each petal starts at this stage. At stage 9 only the dorsal petal is visible; it is still folded and completely covers the lateral and ventral petals. The dorsal petal is as long as the sepals. Veins of each petal are evident at this stage. At stage 10 the dorsal petal is still folded and covers the lateral and ventral petals; however, its size has increased and now the exposed part is as long as the sepals. At stage 11 the dorsal petal is twice the size of the sepals. It is still folded over the lateral and ventral petals, but both the lateral and ventral petals are visible. At stage 12 the three types of petals are completely exposed but their final position in the mature flower is not yet reached. At stage 13 all petals are fully developed and the final disposition of each petal in the flower is established. Petal growth was measured as length and width of each petal during the 7-13 developmental stages.  6.2.3 Scanning electron microscopy (SEM) and light microscopy Full open mature flowers at anthesis of each species were analyzed using a Hitachi S2600N scanning electron microscopy (SEM) at 10 to 12 Kv of acceleration voltage or from fresh flowers preserved on ethanol 70% and analyzed with a light microcope. Some species were analyzed from herbarium specimens, re-hydrated and preserved in ethanol 70%. The distribution of the cell types was analyzed on each type of petal (dorsal, lateral and ventral) in both adaxial and abaxial sides. The epidermal types were classified following a previous study within Leguminosae (Ojeda et al., 2009).  171  6.2.4 RNA extraction and RT-PCR Expression of CYCLOIDEA homologues was studied in five species: two from section Lotus, one from the section Pedrosia and two from the section Rhyncholotus. The expression patterns have been previously reported in L. japonicus, but only in early flower development (stage 0-7) and in mature flowers (Feng et al., 2006b). L. filicaulis is a closely related species of L. japonicus and these two species were included as a comparison of previous expression patterns reported in the model legumes L. japonicus. I selected L. sessilifolius from the section Pedrosia as it has been identified as one of the closely related species to the bird-pollinated species (see Chapter 4). L. sessilifolius has a typical melittophilous flower morphology and flower size and petal arrangement similar to L. japonicus (Fig. 6.2H). Finally, two species with an ornithophilous flower morphology, L. berthelotii and L. maculatus, were included from the section Rhyncholotus (Fig. 6.2A, C). Tissue from each petal type and from three different developmental stages (8, 10 and 13) of each species were extracted using the Pure Link TM Plant RNA Reagent form Invitrogen following manufacturer’s protocol. RNA was treated with DNAse and it was visualized on an agarose gel (2%) and its quantity was measured using a Nanodrop. The RNA was converted to cDNA using the RevertAid TM H Minus First Strand cDNA Synthesis Kit from Fermentas according to manufacturer’s protocol. Genomic contamination was assessed using the intron in LjCYC2 with SL1716/SL1717 (Feng et al., 2006b). Endogenous expression of Lotus japonicus Ubiquitin (LjUbi) was examined with LjUbif/R as a control of cDNA quantity across the samples (Feng et al., 2006b). Gene specific semi-quantitative reverse transcriptase (RT-PCR) was performed on two CYCLOIDEA homologues, LjCYC2 using a new primer LC2.1F (5’ TCCCTTTCAGCTCAAGCCCTTACCC 3’) and LEGCYCR (5’ TCCCTTGCTCTTGCTCTTGC 3’), and LjCYC3 using LEGCYCF1 (5’ TCAGGGSYTGAGGGACCG 3’) and CYC3.1R (5’ 172  CCTGCTTCCTTATTAGGGATTGC 3’) using 95°C for 2 min, 30 cycles of 94°C 45 s, 50-57°C for 1 min and 72°C for 2 min with a final step of 72°C for 5 min. All primers for these CYCLOIDEA homologues were designed in this study, except LEGCYCF and LEGCYCR which have been used previously (Citerne et al., 2003; Citerne et al., 2006; Ree et al., 2004).  6.3 Results 6.3.1 Petal micromorphology of flowers at anthesis within Loteae and the lateralization of petals in section Rhyncholotus and Pedrosia I recorded three major epidermal types within Loteae.Their distribution is mainly restricted to specific petals (Table 6.2). The dorsal petal is characterized by papillose conical cells with striations (PCS). The lateral petal is mainly characterized by tabular rugose cells with striations (TRS) or by a combination of TRS and PCS. In general TRS is located at the base of the petal and PCS, when present, is situated at the tip of the petal and usually on the exposed side (abaxial side) of the petal. Tabular flat cells with striations (TFS) characterize the ventral petal. No PCS were recorded on the ventral petal in any of the species analyzed in this study (Fig. 6.4AF). The majority of the species analyzed in the Loteae and specifically in the section Pedrosia have this distribution of epidermal types (Table 6.2). The four bird-pollinated species have in general an increase in area covered by TRS in the dorsal and ventral petals. The dorsal petal completely lacks papillose conical cells (PCS) and only a small section of the abaxial side of the lateral petal has PCS (Fig. 6.4G-M). The abaxial side of the dorsal petal has elongated non-differentiated cells (ND) which are characteristic of early developmental stages before differentiation. These four species also have trichomes on the abaxial side of the dorsal and lateral petals (Fig. 6.4G). This epidermal distribution suggests a lateralization (increased amount of TRS and absence of PCS) in the dorsal and ventral petals. 173  I also observed a lateralization of the abaxial side of the dorsal petal in species closely related to the bird-pollinated clade, but all these species still have PCS on the adaxial side in the dorsal petal. Six of these species also have trichomes but only on the abaxial side of the dorsal petal (Table 6.3).  6.3.2 Expression patterns of CYCLOIDEA homologues during flower development All three CYCLOIDEA homologues analyzed are expressed asymmetrically within a dorsiventral axis in the flower (dorsal and lateral petals) in all five species studied. LjCYC2 is expressed early during flower development in three species, L. japonicus, L. filicaulis and L. sessilifolius, with a similar flower morphology and similar epidermal types on each petal. The expression of this gene is still observed at late developmental stages but it tends to reduce as the flower completes its development. On the other hand, LjCYC2 expression was not detected in early developmental stages of the two bird-pollinated species analyzed. The earliest expression of LjCYC2 in these two species was observed at stage 10. Expression of LjCYC3 was not detected in the early stage of development in any of the species from section Lotus (L. japonicus and L. filicaulis). The earliest expression of this gene was not detected until stage 10 in L. japonicus and later on L. filicaulis (stage 13). Contrarily, the expression of this gene was detected early on the two bird-pollinated species and also in its closely relative from section Pedrosia, L sessilifolius. The expression continued more or less similar at late develomental stages in all three species (Fig. 6.5). No major changes on LjCYC1 expression were observed between the bee and bird-pollinated species, and only minor discrepancies within each pollination type (Fig. 6.5).  174  6.4 Discussion 6.4.1 Lateralization of the dorsal petal in the bird-pollinated species and the adaptive value of papillose conical cells in melittophilous species The evolutionary transition from melittophilous ancestors to ornithophilous species within Macaronesian Lotus required several floral modifications. Flower symmetry was not altered during the transition (both syndromes have zygomorphic flowers), but the size and the role of each petal in pollinator attraction and pollen placement is modified. The SEM survey also indicates that the epidermal micromorphology in the four birdpollinated species has been modified in comparison with the rest of the species within the tribe Loteae and also in comparison with its closely related species within Pedrosia (Fig. 6.4). These results indicate that the dorsal and ventral petals in the ornithophilous species have suffered a lateralization, an increase of the presence of TRS instead of the typical epidermal types that characterizes each petal (Feng et al., 2006b). This lateralization is particularly evident in the dorsal petal where papillose conical cells are absent. Most of the species from the tribe Loteae and the section Pedrosia analyzed in this study have a distribution of epidermal types reported before (Table 6.2). Papillose conical cells are mainly distributed in the dorsal petal and tabular flat cells with striations (TFS) are located exclusively on the ventral petal (Feng et al., 2006b; Ojeda et al., 2009). These differences of epidermal types between the ornothophilous species and their closest melittophilous relatives within Pedrosia, suggest that the modification of epidermal types must have an adaptive value. Papillose conical cells have an important role in pollinator attraction by increasing the light that is reflected from the flower (Comba et al., 2000; Glover and Martin, 1998, 2002; Martin and Glover, 2007; Noda et al., 1994). Conical cells can also provide tactile cues for the pollinators (Kevan and Lane, 1985a) and it has been shown that this epidermal type 175  also provides an aids to the bee´s grip for flower handling, thus increasing foraging efficiency (Whitney et al., 2009a). The above evidence together with the distribution of papillose conical cells in the adaxial side of the dorsal petal and the abaxial side at the tips of the lateral petals (Table 6.1) suggest that this epidermal type may have a dual role within the flower depending of its location, in pollinator attraction when located in the dorsal petal and in aiding petal handling and foraging efficiency when located on the abaxial side of the lateral petal (as the lateral petal in many papilionoids works as a landing platform for insects). Additionally, papillose cells tend to be mainly distributed on the exposed sides of the petals within the Leguminosae and have been reported on the ventral petal when this petal type has an exposed position within the flower (Ojeda et al., 2009). In melittophilous species the dorsal petal has a major role in attraction, as it is the main area exposed when a pollinator approaches the flower (Fig. 6.3), while lateral petals serve as a landing platform (Proctor et al., 1996). On the other hand, due to its position in the flower and its orientation, the dorsal petal in the bird-pollinated does not have a major role on pollination attraction. It has a reduced size and it is bent backwards. As a consequence the dorsal petal is no longer the main exposed part of the flower functioning in pollinator attraction (Olesen, 1985). Its reflexed position within the flower also allows the insertion of the beak when a bird is foraging for nectar. Additionally, this petal type completely lacks PCS cells. The adaptive value of PCS in this side of the petal is reduced and this epidermal type has been replaced by TRS. Papillose conical cells were only observed in a reduced area of the lateral petal. This cell type is located in a more exposed location, perhaps to guide birds to the location of the nectar at the base of the flower (Fig. 6.4, arrows).  176  Loss of papillose conical cells has also been reported during a transition of pollination syndromes from insect to wind pollination in Thalictrum. In this group, conical cells are distributed in petaloid organs (either sepals or stamen) where these organs have a role in pollination attraction. Wind pollination is derived within this group and conical cells are absent on the sepals and stamen filaments (Di Stilio et al., 2009). My analysis also suggests that this modification in petal micromorpholgy within Pedrosia s.l. is gradual. I found a slight lateralization of the dorsal petal within a group of four species (L. sessilifolius, L. arinagensis, L. kunkelii and L. mascaensis) refered to as the L. sessilifolius group (Sandral et al., 2006) (Table 6.3). This group belongs to same clade as the bird-pollinated species and is the most closely related group with a bee pollination syndrome. However, the lateralization is confined only to the abaxial side, the least exposed area side of the dorsal petal. This gradual evolutionary modification of epidermal types is also supported by the distribution of trichomes. Three of the above mentioned species from the L. sessilifolius group also have trichomes on the dorsal petal. It is worth mentioning that this lateralization is not exlusive to this clade, as other species within the section Pedrosia also have a lateralization of the dorsal petal and trichomes on the abaxial side of the dorsal petal (Table 6.3).  6.4.2 Late expression of LjCYC2 is associated with modification of epidermal types and the evolution of ornithophily CYCLOIDEA is a transcription factor required for zygomorphy establishment in Anthirrhinum majus (Luo et al., 1996). Orthologues of this gene have been identified and isolated in other asterids (Gerbera, Linaria, Gesneriaceae, and Senecio) (Broholm et al., 2008; Cubas and Vincent, 1999; Kim et al., 2008a; Zhou et al., 2008) and in rosids (Iberis, Lotus, Lupinus and 177  Pisum) (Busch and Zachgo, 2007; Citerne et al., 2006; Feng et al., 2006b; Wang et al., 2008) where they play a role in determination of flower symmetry. Modifications of expression patterns in some of these orthologues have been associated with modification of flower symmetry (Broholm et al., 2008; Busch and Zachgo, 2007; Citerne et al., 2006; Zhou et al., 2008) and there is growing evidence that suggests that CYCLOIDEA-like genes can have several effects on flower development, depending of the timing of their expression. Four homologues, LjCYC1, 2, 3 and 5, have been identified in the legume family (Citerne et al., 2003), and at least two of these genes, LjCYC2 and LjCYC3, seem to play a major role in flower symmetry and petal identity in the legumes examined so far (Citerne et al., 2006; Cronk, 2006b; Feng et al., 2006b; Ree et al., 2004; Wang et al., 2008). LjCYC2 is expressed very early during flower development in L. japonicus (before stage 7) and its expression specifies the dorsal-ventral symmetry of the flower. This gene is also required for dorsal petal identity and the development of papillose conical cells (PCS) in this petal. Antisense transgenic plants of L. japonicus lack PCS on the dorsal petal, while plants that overexpress this gene develop PCS on all three types of petals, thus dorsalizing the flower (Feng et al., 2006b). PCS are formed early during flower development and I observed early expression of this gene in all three species analyzed that have PCS on the dorsal petal (Fig. 6.5). In contrast, the two bird-pollinated species analyzed do not develop PCS on the dorsal petal and the absence of this epidermal type is associated with an absence of an early LjCYC2 expression (Fig. 6.5). Late expression of CYCLOIDEA-like genes seems to have additional effects ofnflower development. For example, a homologue of CYCLOIDEA is expressed late during flower development in Iberis amara and Oreocharis benthamani and this late expression represses growth on the adaxial side of the flower (dorsal petal) (Busch and Zachgo, 2007; Du and Wang, 2008). CYCLOIDEA-like genes can also repress or abort other organs within the flower when 178  expressed late during development. There is evidence that OpdCYC1, a CYC-like gene from Opithandra ghushanensis (Genneriaceae), is expressed early during development in all five stamen primordia, but late during development its expression is concentrated in the dorsal and ventral stamens that are reduced or aborted in this species (Song et al., 2009a). Late expression of LjCYC2 in the dorsal petal of the bird-pollinated species may also reduces petal growth. I observed that the growth of the dorsal petal in the ornithophilous species is reduced in comparison with the other two petals (Fig. 6.6A) and this contrasts with the growth pattern that I observed in the melittophilous species, which showed a similar rate of growth of all petals (Fig. 6.6B and C) with a tendency to have a reduction of LjCYC2 expression in late developmental stages (Fig. 6.5). Therefore, late expression of LjCYC2 in the bird-pollinated species may have two effects on flower development in this group; a lack of PCS differentiation, and a reduction of petal growth, both features that are flower modifications that ocurred during the evolutionary transition from mellitophily to ornitophily in this group.  6.4.3 Early expression of LjCYC3 is associated with a lateralization of the flower and suggests a molecular pre-adaptation to bird pollination An additional gene, LjCYC3 specifies lateral petal identity and the formation of TRS on this petal. Functional analyses in L. japonicus have shown that an overexpression of LjCYC3 increases the presence of TRS in all three types of petal and causes a lateralization, but its down regulation ventralizes the lateral petal, with an absence of TRS on this petal (Feng et al., 2006b). Here I found that LjCYC3 expression is detected late during flower development in L. japonicus (Fig. 6.5), confirming previous studies that also indicate that this gene is not detected during early developmental stages (0-6) using in situ hybridization (Feng et al., 2006b). The early expression of this gene observed in both ornithophilous species and its closely related species L. sessilifolius 179  is associated with an increase of TRS on the dorsal petal, thus supporting the idea that all these three species have a lateralized dorsal petal (Fig. 6.5). It has been suggested that LjCYC2 and LjCYC3 interact during flower development. Early expression of LjCYC2, combined with a lack of early expression of LjCYC3 in L. japonicus are necessary for the differentiation of PCS on the dorsal and TRS on the lateral petal of this species (Da Luo, unpubl.). The modifications of the epidermal types and petal identities in the ornithophilous species seem to be explained by a combination of two heterochronic processes during the expression of CYCLOIDEA homologues. A delayed (or post-displacement) expression of LjCYC2 combined with a precocious (pre-displacement) of LjCYC3 expression during flower development. My analysis does not indicate a change of heterotopy, as reported in other flower modifications, such as that reported in Cadia purpurea (Citerne et al., 2006). The distribution of epidermal types in L sessilifolius is also explained by these two mechanisms. This species has a slightly less lateralized dorsal petal. In constrast to the ornithophilous species, this species has an early expression of both genes, LjCYC2 and LjCYC3 and the lateralization of the flower is weaker in comparison with the ornithophilous species. Given that this species is the closest relative to the four bird-pollinated species (see Chapter 4), it is possible that this evolutionary change of epidermal types and flower morphology required different steps of heterochronic modifications of expression of these genes, first a precocious expression of the lateral identity gene LjCYC3, followed by a down regulation of the dorsal petal identity gene. Taken together my results indicate that the clade where the bird pollination syndrome evolved in Pedrosia may have been facilitated by a molecular pre-adaptation of these two CYCLOIDEA homologues.  180  Table 6.1 List of samples collected for the analysis of epidermal types in Loteae Taxon Acmispon Section Anisolotus (Benh.) D.D. Sokoloff A. americanus (Nutt.) Rydb. Anthyllis A. hermannieae L. Coronilla C. valentina L. C. varia L. Hosackia H. chihuahuana S. Watson Lotus Section Bonjeana (Rchb.) D.D. Sokoloff L. hirsutus L. Section Chamaelotus Kramina and D.D. Sokoloff L. glinoides Del. Section Erythrolotus Brand L. conimbrensis Brot. Section Heineckenia Webb & Berth. L. arabicus L. L. gebelia Vent. Section Krokeria (Moench) Ser. L. edulis L. Section Lotea (Medik.) DC. L. halophilus Boiss. & Spruner L. weilleri Maire Section Lotus L. burttii Borsos L. corniculatus L. L. filicaulis Durieu L. japonicus Gifu B129 Section Pedrosia (Lowe) Christ L. arborescens Lowe ex Cout. L. arenarius Brot. L. arinagensis Brawm. L. argyrodes R.P Murray L. assakensis Brand L. azoricus P. W. Ball L. brunneri Webb in Hooker L. callis-viridis Bramwell & D.H. Davis L. campylocladus Webb & Berthel. L. creticus L. L. dumetorum Webb ex R. P. Murray L. emeroides R. P. Murray L. eriosolen (Maire) Mader & Podlech L. genistoides Webb & Berthel. L. glaucus Sol. L. hillebrandii Christ L. holosericeus Webb & Berthel. L. jacobaeus L. L. jolyi Battand. L. kunkelii (Esteve) Bramwell & D. H. Davis  Collection information  Voucher/Herbarium  Cult. UBC PI 215232  Ojeda 240/UBC  UBCBG, No. 035419-0389-2000  Ojeda 138/UBC  UBCBG, without number UBC campus, Vancouver  Ojeda 33/UBC Ojeda 39/UBC  Cult. UBC PI 18085  Ojeda 79/UBC  UBCBG, No. 032962-0447-1996  Ojeda 58  Cult. PI 246736  -  Cult. PI 238334  -  Cult. UBC PI 214109 Cult. PI 464685  Ojeda 151/UBC  Cult. UBC PI 244281  Ojeda 152/UBC  Cult. UBC PI 238336 Cult. UBC PI 631729  Ojeda 153/UBC Ojeda 154/UBC  Cul. UBC seeds, Miyasaki University UBC campus, Vancouver Cult. UBC Cult. UBC  Ojeda 72  Cult. JCVC # 164/06 Cult. UBC PI 631780 Cult. JCVC # 49/04 Cult. JCVC # 5435/UDH/07 Voucher  Ojeda 180/UBC Ojeda 78/UBC Ojeda 178/UBC Ojeda 189/UBC Jury & Upson 20510/RNG ORT # 36336 Ojeda 181/UBC Ojeda 177/UBC Ojeda 206/UBC Ojeda 188/UBC Ojeda 213/UBC Ojeda 228/UBC Ojeda 209/UBC  Voucher Cult. JCVC # 514B/07 Cult. JCVC # 145/04 Carretera al Teide, T Cult. JCVC # 64/07 Mirador Jardina, T Teno, T. Epina, G Cult.UBC # PI 631784, Tiznir, Morocco Cañon del Jierro, GC Cult. JCVC # 223/B/07 Cumbre nueva a Fuencaliente, P Cult. JCVC # 334/02 Cult. JCVC 183/06 Voucher Cult. JCVC # 217/07  Ojeda 27 Ojeda 71 Ojeda 70  Ojeda 174/UBC Ojeda 187/UBC Ojeda 232/UBC Ojeda 201/UBC Ojeda 179/UBC S.L. Jury & Upson 20503/RNG Ojeda 176/UBC  181  Taxon L. lancerottensis Webb & Berthel. L. latifolius Brand L. leptophyllus (Lowe) K. Larsen L. macranthus Lowe L. maroccanus Ball L. mascaënsis Burchard L. pseudocreticus Maire, Weiller & Wilczek L. purpureus Webb L. salvagensis R.P. Murray L. sessilifolius D.C. L. sessilifolius D.C. L. spartioides Webb & Berthel. L. tenellus (R. Lowe) Sandral, Santos & D.D. Sokoloff Section Rhyncholotus (Monod) D.D. Sokoloff L. berthelotii Masf. L. eremiticus A. Santos L. maculatus Breitf. L. pyranthus P. Perez Ornithopus O. compressus L. Scopiurus S. sulcata L.  Collection information Voucher Cult. JCVC 159/06 Puente de Silva, GC Voucher Voucher Cult. JCVC # 442/99 Voucher  Voucher/Herbarium ORT # 37824 Ojeda 183/UBC Ojeda 171/UBC ORT # 36675 Jury 14471/RNG Ojeda 200/UBC Davies 53484/RNG  Cult. JCVC # 167/06 Voucher Guimar Poligono industrial, T H Tamadaba, GC Playa de los Roques, T  Ojeda 184/UBC ORT # 35128 Ojeda 224/UBC  Cult. UBC LB-08 Cult. JCVC 235/07 Cult. JAO # 100/06 Cult UBC LM-03 Cult. JCVC # 187/07 Cult. JCVC # 210/99 Cult. JAO # 124-01  Ojeda 238/UBC Ojeda 185/UBC Ojeda 239/UBC Ojeda 186/UBC Ojeda 175/UBC Ojeda 226/UBC  Barranco Madera, P  Ojeda 158/UBC  Valle de Masca, T  Ojeda 160/UBC  Ojeda 217/UBC Ojeda 215/UBC  182  Table 6.2 Distribution of epidermal types on species analyzed. PCS= papillose conical cells, TRS= tabular rugose cells with striations, TFS= tabular flat cells. Epidermal types separated by a dash indicate that two epidermal types where observed on the same petal, s= stomata, t= trichomes Taxon Acmispon Section Anisolotus A. americanus Anthyllis A. hermannie Coronilla C. valentine C. varia Hosackia H. chihuahuana Lotus Section Bonjeana L. hirsutus Section Chamaelotus L. glinoides Section Erythrolotus L. conimbrensis Section Heineckenia L. arabicus L. gebelia Section Krokeria L. edulis Section Lotea L. halophilus L. weilleri Section Lotus L. burttii L. corniculatus L. filicaulis L. japonicus Gifu B129 Section Pedrosia L. arborescens L. arenarius L. arinagensis L. argyrodes L. azoricus L. assakensis L. brunneri L. callis-viridis L. campylocladus L. creticus L. dumetorum L. emeroides L. eriosolen L. glaucus L. genistoides L. holosericeus L. hillebrandii L. jacobaeus  Dorsal petal abaxial adaxial  Lateral petal abaxial adaxial  Ventral petal abaxial adaxial  TRSs  TRS  TRS  TRS  TFS  TFS  PCS  PCS  TRS  TRS  TFS  TFS  PCS PCS  PCS PCS  PCS PCS  PCS TRS  TRS/TFS TFS  TFS TFS  PCS  PCS  PCS  TRS  TRS/TFS  TFS  PCS  PCS  TRS  TRS  TFS  TFS  TRS  PCS  TRS/PCS  TRS  TFS  TFS  PCS  PCS  TRS/PCS  TRS  TFS  TFS  TRS PCS  PCS PCS  TRS/PCS TRS/PCS  TRS TRS  TFS TFS  TFS TFS  TRS  PCS  TRS/PCS  TRS  TFS  TFS  TRS PCS  PCS PCS  TRS/PCS TRS/PCS  TRS TRS  TFS TFS  TFS TFS  PCS PCS PCS PCS  PCS PCS PCS PCS  TRS TRS TRS TRS  TRS TRS TRS TRS  TFS TFS TFS TFS  TFS TFS TFS TFS  PCS PCS TRS TRS/PCS TRS/PCS PCS PCS PCS PCS PCS TRS PCS TRS PCS PCS PCS PCS PCS  PCS PCS TRS TRS PCS PCS PCS PCS PCS PCS PCS PCS TRS PCS PCS PCS PCS PCS  PCS PCS TRS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS  TRS TRS TRS/PCS  TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS  TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS  TRS/PCS TRS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS  183  Taxon L. jolyi L. kunkelii L. lancerottensis L. latifolius L. leptophyllus L. macranthus L. maroccanus L. mascaënsis L. pseudocreticus L. purpureus L. salvagensis L. sessilifolius subsp. sessilifolius L. sessilifolius subsp. villosissimus L. spartioides L. tenellus Section Rhyncholotus L. berthelotii L. eremiticus L. maculatus L. pyranthus Ornithopus O. compressus Scopiurus S. sulcata  Dorsal petal abaxial adaxial TRS PCS TRSt PCS PCS PCS PCS PCS TRS PCS TRS PCS TRS PCS TRSt PCS PCS PCS PCS PCS PCS PCS TRS PCS  Lateral petal abaxial adaxial TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS TRS/PCS  Ventral petal abaxial adaxial TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS TFS  TRSt  PCS  TRS/PCS  TRS/PCS  TFS  TFS  TRS TRS  PCS PCS  TRS/PCS TRS/PCS  TRS/PCS TRS/PCS  TFS TFS  TFS TFS  NDt NDt NDt NDt  TRS TRS TRS TRS  TRSt/PCS TRSt/PCS TRSt/PCS TRSt/PCS  TRS TRS TRS TRS  TRS TRS TRS TRS  TRS TRS TFS TRS  PCS  PCS  PCS  TRS/PCS  TFS  TFS  TRSS  PCS  TRS  TRS  TFS  TFS  184  Table 6.3 Classification of the levels of lateralization (meassured by the presence of TRS) observed in Rhyncholotus and Pedrosia species according with the distribution of epidermal types and trichomes. * Species reported with trichomes on the dorsal petal (Sandral et al., 2006), but not confirmed in this study. L. chazalei and L. loweanus were not analyzed. The specimen of L. assakensis analyzed did not have trichomes.  Syndrome  Section Rhyncholotus Ornithophilous species  Groups  Species  Epidermal type  Lateralization  Trichomes PCS  Section Pedrosia Melittophilous species Closely related species of Rhyncholotus  L. berthelotii L. maculatus L. pyranthus L. eremiticus  -Modified TRS and increased presence of this epidermal type in dorsal and ventral petal -Strong lateralization of dorsal and ventral petals -Present on dorsal and lateral petals -Absent on dorsal petal  Species within the same clade of Rhyncholotus L. argyrodes L. azoricus L. arinagensis L. sessilifolius subsp. sessilifolius  Species from a different clade of Rhyncholotus L. spartioides L. tenellus L. leptophyllus L. dumetorum  Species from a different clade of Rhyncholotus *L. assakensis *L. loweanus *L. chazalei -Increased presence of TRS on the abaxial side of the dorsal petal  - Increased presence of TRS on the abaxial side of the dorsal petal  - Increased presence of TRS on the abaxial side of the dorsal petal  -Slight lateralization of dorsal petal  -Slight lateralization of dorsal petal  -Slight lateralization of dorsal petal  -Present on dorsal petal  -No trichomes  -No trichomes  -PCS on dorsal petal  -PCS on dorsal petal  -PCS on dorsal petal  L. mascaënsis L. kunkelii L. sessilifolius subsp. villosissimus  185  Figure 6.1 Developmental stages of flower in Lotus japonicus  A  Stage 8  Stage 11  Stage 9  Stage 12  Stage 10  Stage13  186  Figure 6.2 Members of Lotus section Rhyncholotus with an ornithophilous pollination syndrome. A, Lotus maculatus, B, L. pyranthus, C, L. berthelotii, and D, L. eremiticus; and four representative species of section Pedrosia, E, L. arenarius, F, L. latifolius, G, L. jacobaeus and H, L. sessilifolius subsp. sessilifolius with a bee-pollination syndrome. a=dorsal petal, b = lateral petal and c=ventral petal. Photo credits: A, C, E, H from I. Ojeda and B, D, F, and G from F. Oliva-Tejera.  b  a b c  c  a  A  B  c b  b  c a  a  C  D a  a  b c  b c  H  G  a  a b  c  b c  E  F  187  Figure 6.3 Zygomorphic flowers of A, Lotus japonicus. B, these flowers have three types of petals, a) dorsal petal with a bilateral symmetry, b) two asymmetrical lateral petals, and c) two asymmetrical ventral petals. All petals were separated and flattened. The base of the dorsal petals in both species was separated from the rest of the petal. C, Lotus berthelotii with a bird-pollination syndrome. D, size comparison between L. berthelotii and L. japonicus. E, hypothetical mechanism by which a bird seeks for nectar and pollen is placed either on the top of the head or in the throat (according to Olsen, 1985). E, Bombus canariensis canariensis one of the main insect pollinators in the Canary Islands foraging on L. hillebrandii from El Hierro.  a  a  c b  b  b  c A  B  D  E  c  a  b  C  F  188  Figure 6.4 Major epidermal types recorded in Loteae. A and B, papillose conical cells (PCS), C and D, tabular rugose cells with striations (TRS) and E an F, tabular flat cells with striation (TFS) in Lotus japonicus. G, non-differentiated cells with trichomes, H, tabular rugose cells with striations, I, papillose conical cells with striations, H, J-M tabular rugose cells with striations in L. berthelotii (TRS). Arrows indicate the localization of papillose conical cells in the lateral petals. Scale bars: 100 μm, G and L. 50 μm = A, H, C, D, F, K and M. 25 μm = I and J.  Dorsal A  B  G  C  D  I  E  F  L  H  Lateral J  K  Ventral  Abaxial  Adaxial  M Abaxial  Adaxial  189  Figure 6.5 Expression patterns of LjCYC1, LjCYC2 and LjCYC3 in A) Lotus japonicus, B) L. filicaulis both from section Lotus, C) L. sessilifolius from section Pedrosia. All these three species have a flower morphology adapted to bee pollination and have similar petal size and shape. D) L. berthelotii and E) L. maculatus from section Rhyncholotus have a flower morphology adapted to ornitophily by opportunistic passerines. The base of the dorsal petal has been detached from the rest of the petal. dp= dorsal petal, lp= lateral petal and vp=ventral petal.  190  A  dp  Stage 13  Stage 10  Stage 8  dp lp vp  dp lp vp  dp lp  vp  dp lp vp  dp lp vp  dp lp  vp  dp lp vp  dp lp vp  dp lp  vp  Stage 13  Stage 10  dp lp vp  dp lp vp  dp lp  vp  dp lp vp  dp lp vp  dp lp  vp  LjCyc1  Section Lotus  LjCyc2  dp  lp  LjUbiquitin  B dp lp vp  Section Pedrosia  LjCyc3  vp  lp  LjCyc1  lp  LjCyc2 LjCyc3  vp  vp  LjUbiquitin  LjCyc1  C  LjCyc2  dp lp vp  LjCyc3 LjUbiquitin  Melittophilous  dp  Section Rhyncholotus  D  vp  Stage 8  lp LjCyc1  dp  LjCyc2 LjCyc3  E  lp  dp lp vp  lp  LjUbiquitin  LjCyc1  vp  vp  LjCyc2 LjCyc3  Ornithophilous  LjUbiquitin  191  Figure 6.6 Petal gowth meassured as the length/width ratio during different stages of flower development in A) L. berthelotii, B) L. sesilifolius and C) L. japonicus. Mean values with standard deviations. (n=3-5 flowers on each developmental stage).  Length / width ratio  A  Length / width ratio  B  Length / width ratio  C  Flower developmental stages 192  6.5 Bibliography Allan, G.J., Francisco-Ortega, J., Santos-Guerra, A., Boerner, E., Zimmer, E., 2004. Molecular phylogenetic evidence for the geographic origin and classification of Canary Island Lotus (Fabaceae: Loteae). Molecular Phylogenetics and Evolution 32, 123-138. Bradshaw, H.D.J., Otto, K.G., Frewen, B.E., McKay, J.K., Schemske, D.W., 1998. Quantitative trait loci affecting differences in floral morphology between two species of monkeyflower (Mimulus). Genetics 149, 367-382. Bradshaw, H.D.J., Schemske, D.W., 2003. Allele substitution at a flower color locus produces a pollinator shift in monkeyflowers. Nature 426, 176-178. Broholm, S.K., Tahtiharju, S., Laitinen, R.A., Albert, V.A., Teeri, T.H., Elomaa, P., 2008. A TCP domain transcription factor controls flower type specification along the radial axis of the Gerbera (Asteraceae) inflorescence. Proceedings of the National Academy of Sciences of the United States of America. Busch, A., Zachgo, S., 2007. Control of corolla monosymmetry in the Brassicaceae Iberis amara. Proceedings of the National Academy of Sciences of the United States of America 104, 16714-16719. Carracedo, J.C., Pérez, F.J., Ancochea, E., J., M., Hernán, F., C.R., C., Casillas, R., Rodriguez, E., Ahijado, A., 2002. Cenozoic volcanism II: The Canary Islands. In: Gibbons, W., Moreno, T. (Eds.), The Geology of Spain The Geological Society Publishing House. , London, pp. 439-472. Citerne, H.L., Luo, D., Pennington, R.T., Coen, E., Cronk, Q.C.B., 2003. A phylogenomic investigation of CYCLOIDEA-like TCP genes in the Leguminosae. Plant Physiology 131, 1042-1053.  193  Citerne, H.L., Pennington, R.T., Cronk, Q.C., 2006. An apparent reversal in floral symmetry in the legume Cadia is a homeotic transformation. Proceedings of the National Academy of Sciences United States of America 103, 12017-12020. Comba, L., Corbet, S., Hunt, H., Outram, S., Parker, J.S., Glover, B.J., 2000. The roles of genes influencing the corolla in pollination of Antirrhinum majus. Plant, Cell and Environment 23, 639-647. Cronk, Q.C., 2006. Legume flowers bear fruit. . Proceedings of the National Academy of Sciences United States of America 103, 4801-4802. Cronk, Q.C., Ojeda, I., 2008. Bird-pollinated flowers in an evolutionary and molecular context. Journal of Experimental Botany 59, 715-727. Cubas, P., Vincent, C., 1999. An epigenetic mutation responsible for natural variation in floral symmetry. Science 401, 157. Degtjareva, G.V., Kramina, T.E., Sokoloff, D.D., Samigullin, T.H., Valiejo-Roman, C.M., Antonov, A.S., 2006. Phylogeny of the genus Lotus (Leguminosae, Loteae): evidence from nrITS sequences and morphology. Canadian Journal of Botany 84, 813-830. Di Stilio, V.S., Martin, C., Schulfer, A.F., Connelly, C.F., 2009. An ortholog of MIXTA-like2 controls epidermal cell shape in flowers of Thalictrum. New Phytologist 183, 718-728. Dong, Z.C., Zhao, Z., Liu, C.W., Luo, J.H., Yang, J., Huang, W.H., Hu, X.H., Wang, T.L., Luo, D., 2005. Floral patterning in Lotus japonicus. Plant Physiology 137, 1272-1282. Du, Z.Y., Wang, Y.Z., 2008. Significance of RT-PCR expression patterns of CYC-like genes in Oreocharis benthamii (Gesneriaceae). Journal of Systematics and Evolution 46, 23-31. Feng, X., Zhao, Z., Tian, Z., Xu, S., Luo, Y., Ca, Z., Wang, Y., Yang, J., Wang, Z., Weng, L., Chen, J., Zheng, J., Guo, X., Luo, J., Sato, S., Tabata, S., Ma, W., Cao, X., Hu, X., Sun, C., Luo, D., 2006. Control of petal shape and flower zygomorphy in Lotus japonicus. 194  Proceedings of the National Academy of Sciences of the United States of America 103, 4970-4975. Galliot, C., Hoballah, M.E., Kuhlemeier, C., Stuurman, J., 2006. Genetics of flower size and nectar volume in Petunia pollination syndromes. Planta 225, 203-212. Glover, B.J., Martin, C., 1998. The role of petal shape cell and pigmentation in pollination success in Antirrhinum majus. Heredity 80, 778-784. Glover, B.J., Martin, C., 2002. Evolution of adaptive petal cell morphology. In: Cronk, Q.C.B., Bateman, R.M., Hawkins, J.A. (Eds.), Developmental genetics and plant evolution. Taylor & Francis, London, pp. 160-172. Grant, K.A., Grant, V., 1968. Hummingbirds and their flowers. Columbia University Press, New York and London. Hoballah, M.E., Gubitz, T., Stuurman, J., Broger, L., Barone, M., Mandel, T., Dell'Olivo, A., Arnold, M., Kuhlemeier, C., 2007. Single gene-mediated shift in pollinator attraction in Petunia. Plant Cell 19, 779-790. Kevan, P.G., Lane, M.A., 1985. Flower petal microtexture is a tactile cue for bees. Proceedings of the National Academy of Sciences of the United States of America 82, 4750-4752. Kim, M., Cui, M.L., Cubas, P., Gillies, A., Lee, K., Chapman, M.A., Abbot, R.J., Coen, E., 2008. Regulatory genes control a key morphological and ecological trait transferred between species. Science 322, 1116-1119. Luo, D., Carpenter, R., Vincent, C., Copsey, L., Coen, E., 1996. Origin of flower symmetry in Anthirrinum. Nature 383, 794-799. Martin, C., Glover, B.J., 2007. Functional aspects of cell patterning in aerial epidermis. Current Opinion in Plant Biology 10, 70-82.  195  Noda, K., Glover, B.J., Linstead, P., Martin, C., 1994. Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor. Nature 369, 661664. Ojeda, I., Francisco-Ortega, J., Cronk, Q.C.B., 2009. Evolution of petal epidermal micromorphology in Leguminosae and its use as a marker of petal identity. Annals of Botany 104, 1099-1110. Olesen, J.M., 1985. The Macaronesian bird flower element and its relation to bird and bee opportunistic. Botanical Journal of the Linnean Society 91, 395-414. Ollerton, J., Cranmer, L., Stelzer, R.J., Sullivan, S., Chittka, L., 2008. Bird pollination of Canary Island endemic plants. Nature Preceedings. Proctor, M., Yeo, P., Lack, A., 1996. The natural history of pollination. Timber Press, Portland. Quattrochio, F., Wing, F.J., van der Woude, K., Souer, E., de Vetten, N., Mol, J.N., Koes, R., 1999. Molecular analysis of the anthocyanin2 gene of petunia and its role in the evolution of flower color. Plant Cell 11, 1433-1444. Ree, R.H., Citerne, H.L., Lavin, M., Cronk, Q.C.B., 2004. Heterogeneous selection on LEGCYC paralogs in relation to flower morphology and the phylogeny of Lupinus (Leguminosae). Molecular Biology and Evolution 21, 321-331. Sandral, G., Remizova, M.V., Sokoloff, D.D., 2006. A taxonomic survey of Lotus section Pedrosia (Leguminosae, Loteae). Wulfenia 13, 97-192. Song, C.F., Lin, Q.B., Liang, R.H., Wang, Y.Z., 2009. Expression of ECE-CYC2 clade genes relating to abortion of both dorsal and ventral stamens in Opithandra (Gesneriaceae). BMC Evolutionary Biology 9, 244. Stelzer, R., 2005. Sammelstrategien bei Hummeln ein Vergleich zwischen Insel und Festlandpopulationen. Fakultät für Biologie. Universität Würzburg. 196  Stuurman, J., Hoballah, M.E., Borges, L., Moore, J., Basten, C., Kuhlemeier, C., 2004. Dissection of floral pollination syndrome in Petunia. Genetics 168, 1585-1599. Wang, Z., Luo, Y., Li, X., Wang, L., Xu, S., Yang, J., Weng, L., Sato, S., Tabata, S., Ambrose, M., Rameau, C., Feng, X., Hu, X., Luo, D., 2008. Genetic control of floral zygomorphy in pea (Pisum sativum L.). Proceedings of the National Academy of Sciences of the United States of America 105, 10414-10419. Whitney, H.M., Chittka, L., Bruce, T.J.M., 2009. Conical epidermal cells allow bees to grip flowers and increase foraging efficiency. Current Biology 19, 948-953. Whittall, J.B., Hodges, S.A., 2007. Pollinator shifts drive increasingly long nectar spurs in columbine flowers. Nature 447, 706-709. Zhang, S., Sandal, N., Polowick, P.L., Stiller, J., Stougaard, J., Fobert, P.R., 2003. Proliferating Floral Organs (Pfo), a Lotus japonicus gene required for specifying floral meristem determinacy and organ identity, encodes an F-box protein. Plant Journal 33, 607-619. Zhou, X.R., Wang, Y.Z., Smith, J.F., Chen, R., 2008. Altered expression patterns of TCP and MYB genes relating to the floral developmental transition from initial zygomorphy to actinomorphy in Bournea (Gesneriaceae). New Phytologist 178, 532-543. Zufall, R.A., Rausher, M.D., 2004. Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428, 847-850.  197  7 Biochemical evolution of bird-pollinated flowers in Macaronesian Lotus (Leguminosae)1  7.1 Introduction Flower colour depends on a combination of many factors, including epidermal surface (Noda et al., 1994), co-pigments, metal ions and pH of the vacuolar content, and the type of pigments produced (Grotewold, 2006; Tanaka et al., 2008). Flavonoids are one of the main pigments in plants and are responsible for the red to blue colours of many plant parts, including flower colours. The flavonoid biosynthetic pathway is well conserved across angiosperms, and it is the metabolic pathway responsible for the production of several of the most common types of plant pigments. Anthocyanins and flavonols are two major pigment types derived from this pathway, the former produce orange/red to violet/purple colours in flowers while the latter can produce very pale-yellow colours. Both types of pigments derive from dihydroflavonols and the type of pigment produced largely depends on the enzymes active downstream of these precursors. Flavonol synthase (FLS) is responsible for producing flavonols, while dihydroflavonol 4reductase (DFR) is the first enzyme for commitment to anthocyanin production. There are three main branches in the anthocyanin pathway and their colour depends mainly on the number of hydroxyl groups on the B-ring. As the number of hydroxyl groups increases, the peak of absorbance shifts from the red side to the blue side of the spectrum (Tanaka et al., 1998); the larger the number of hydroxyl groups the bluer the colour.  1  A version of this chapter will be submitted for publication: Ojeda, I. Santos-Guerra, A., Oliva-  Tejera, F. Alfredo Valido, Jaen Molina, Xinxin, Xue, R. Caujape, J. Marrero, A. and Cronk. Q. C. B. Biochemical evolution of bird-pollinated flowers in Macaronesian Lotus (Leguminosae) 198  Pelargonidin with one hydroxyl group generally produces red colours, cyanidin derivatives with two groups typically give magenta or blue colours, and delphinidin with three hydroxyl groups is mainly responsible for blue colours (Grotewold, 2006; Tanaka et al., 2008). Red/orange flowers are usually associated with bird pollination (ornithophily) and this trait is probably one of the most conspicuous features of this syndrome (Proctor et al., 1996). It is thought that red flowers evolve either to advertise certain features, such as copious nectar, to birds (Grant, 1966), or because it is not attractive to competing insect pollinators (Cronk and Ojeda, 2008); either way, flower colour has an important role in the type of pollinators attracted. Pelargonidin derivatives are often associated with hummingbird pollination, while cyanidin derivatives seem to be associated with pollination by perching birds (Scogin, 1988). Delphinidin derivatives are less frequently found in bird-pollinated species but are often found in blue bee-pollinated flowers. Several transitions in flower colour associated with transitions in pollination syndromes have previously been investigated at the biochemical and molecular level. Bee-bird transitions have been investigated in modifications from blue to red flowers (Rausher, 2008) and yellow to red (Taylor, 1984). Modifications of pigments in association with blue (bee) to red (bird) transitions of flower colour have been previously explored in three genera. Red bird-pollinated species of Ipomoea have derivatives of the pelargonidin branch and the transition from its bee ancestor involved the inactivation of the cyanidin branch (Zufall and Rausher, 2004). Most of the red, hummingbird pollinated, species within this genus have pigments of the pelargonodin branch with no derivatives of other branches of the anthocyanin pathway (Streisfeld and Rausher, 2009b; Zufall, 2003). In Penstemon, another group with hummingbird pollination, red flowers have evolved independently several times and have pigments of the cyanidin or pelargonidin branch, or both (Rausher, 2008; Scogin and Freeman, 1987). 199  In other examples, bird-pollinated flowers have evolved from yellow flowered insectpollinated ancestors. In some Iochroma species, for example, red-flowered hummingbird pollinated species, which have pelargonidin pigments, probably evolved from yellow ancestors that lacked this pigment and produced only carotenoids (Smith, 2006). In Mimulus aurantiacus, there are red-flowered populations associated with hummingbird pollination and yellow mothpollinated populations. It seems that natural selection is maintaining this difference in flower colour (Streisfeld and Rausher, 2009a). Red-flowered populations are able to produce cyanidin and pelargonidin pigments, while yellow-flowered populations completely lack these two pigments (Streisfeld and Kohn, 2005). Red hummingbird pollinated flowers of Aquilegia also produce cyanidin and pelargonidin pigments (Taylor, 1984). In this study my objective is to explore a transition from yellow-flowered species pollinated by insects (bumblebees) to red/orange-flowered pollinated by opportunistic passerines birds in a group of Macaronesian Lotus (Fig. 7.1A and B). Within Macaronesian Lotus, this pollination syndrome is found in a group of four species (the ―rhyncholotus group‖). Additional features associated with this syndrome include flower size, shape, orientation, longevity as well as nectar composition and concentration (Dupont et al., 2004; Olesen, 1985; Ollerton et al., 2009). On the other hand, their most closely related species within section Pedrosia (36 spp) have the standard traits associated with bee pollination found in the genus Lotus. These traits include a horizontal flower orientation, low nectar production (with sucrose as the major sugar) and yellow flowers at anthesis (Fig. 7.1C). Some species within this group alter flower colour after anthesis, in which case old flowers may assume brown, pink, orange, purple or red colours depending on the species (Fig. 7.1B-F). In Lotus the yellow colour is known to be mainly due to carotenoids (Suzuki et al., 2007) but flavonols and anthocyanins are also known to occur (Reymaud and Lussignol, 2005; Suzuki 200  et al., 2008) and are likely to be important modifiers of spectral reflectance. Pigment composition in the bird-pollinated Lotus species has been investigated in only one species, in which a cyanidin-based pigment was identified (Beale et al., 1941). There are no previous analyses of the pigment composition in the Macaronesian group for the bee-pollinated species. Outside of Macaronesian Lotus, pigments have been investigated in the model legume Lotus japonicus (Suzuki et al., 2008).  7.1.1 Objectives of the study My goals in this study are: 1) to identify pigment modifications (anthocyanins and flavonols) associated with the evolutionary transition from yellow bee-pollinated flowers to red/orange bird-pollinated flowers, 2) to determine the biochemical changes associated with postanthesis flower colour modification 3) to determine floral reflectance in relation to pollinator perception in differently coloured Lotus flowers, and 4) to examine the molecular basis of changes in pigmentation by analyzing the relative expression patterns of three biosynthetic enzyme genes between the two pollination syndromes.  7.2 Material and methods 7.2.1 Pigment extraction and composition analysis In order to determine the type of pigments produced in flowers of the species in this group, the petals were separated from the rest of the flower and dried. Pigments were extracted from petal samples (20 mg) with a buffer (MeOH/H20/AcAc) and later treated with HCl 2N at 100°C for 30 min (Shimada et al., 2005). Samples were injected into a LC Agilent 1100 series light chromatograph (LC-MS) with a LC7MSD trap XCT Plus. One species (L. sessilifolius) was  201  analyzed three times to assess variation but the remaining species were analyzed only once as the variation in pigment composition was minimal. Pigments from the samples were identified using retention times of a mixture of six anthocyanin standards (cyanidin, peonidin, pelargonidin, petunidin, delphinidin and malvidin from; Chromadex, and three flavonols, quercetin, kaempferol and isorhamnetin (from SIGMA) prepared and analyzed under similar conditions. A total of 18 species were analyzed representing all flower colours reported in this group. This sample included the four bird-pollinated species, five yellow bee-pollinated species that do not alter flower colour after anthesis and nine beepollinated species that change flower colour after anthesis. My sample also included two species from section Lotus, the model legume L. japonicus and L. filicaulis, a species that changes flower colour after anthesis from yellow to red (Table 7.1). The relative amount of each pigment was calculated from the area under peaks in the MS spectra for each pigment identified using the LC/MSD trap software 5.2. The relative amount of each pigment was used to establish which branches of the anthocyanin pathways were active and to determine pigment pathway alterations during transitions of flower colour.  7.2.2 Measurement of petal reflectance Petal reflectance was measured under field conditions (in situ in the wild and in plants under nursery conditions) with a portable spectrophotometer (Ocean Optics USB2000) at wavelengths from 300 to 700 nm. The three types of petals were measured on the side naturally exposed to pollinator vision (abaxial side of the dorsal petal and adaxial side of lateral and ventral petals). All measurements were taken five times on the same section of the petal (middle part) for each species and the averages of these measurements were used to estimate the reflectance graph for each species. In total 19 species were analyzed, including all four bird-pollinated species, five 202  yellow bee-pollinated species that do not change colour and 10 bee-pollinated yellow species that change colour after anthesis (measured both before and after the colour change) (Table 7.2). The species included all flower colours reported in this group, except pink and brown flowers, which were not available for this study. Reflectance graphs of each species were later classified according to their correspondence to action spectra of the four visual receptors found in pollinating organisms (UV, blue, green and red) (Chittka et al., 1994) and used to determine perceived flower colour by pollinators.  7.2.3 Reconstruction of flower colour change evolution Flower colour modification after anthesis was analyzed using parsimony (DELTRAN) as implemented in MacClade 4.0 (Maddison and Maddison, 2000). Ancestral state reconstruction was carried out using Mesquite (Maddison and Maddison, 2009). Flower colour change was recorded for each species based on observations in the field, from cultivated plants in botanical gardens (Jardin de Aclimatación de la Orotava, Jardin Canario Botánico Viera y Clavijo) and from plants cultivated at the University of British Columbia (UBC) glasshouses. For those species not cultivated or observed in the field, flower colour was obtained from the literature (Bramwell and Bramwell, 2001; Brochman et al., 1997; Jardim and Francisco, 2000; Mader and Podlech, 1989; Monod, 1980; Sandral et al., 2006). Colour change was coded as a binary character (absence and presence of flower colour change) and the evolution of this trait was mapped on a phylogenetic tree of the group based on four gene regions (ITS, Cytochrome B6, trnH-psbA and matK) (see Chapter 4).  203  7.2.4 Gene expression of dihydroflavonol-4-reductase (DFR), anthocyanidin synthase (ANS) and O-methyl transferase (OMT) In order to further explore the activity of the anthocyanin pathway and its different branches in this group during flower colour transitions, I analyzed the expression of three genes using reverse transcriptase PCR (RT-PCR). I extracted RNA from two developmental stages, open mature flower at anthesis (before colour change) and flowers after colour modification (post-change) in each petal separately. RNA was extracted using the Pure LinkTM Plant RNA Reagent from Invitrogen following the manufacturer’s protocol. RNA was treated with DNAse, visualized on an agarose gel (2%) and its concentration measured using a Nanodrop spectrophotometer. RNA was converted to cDNA using the RevertAid TM H Minus First Strand cDNA Synthesis Kit from Fermentas according to the manufacturer’s protocol. Genomic contamination was assessed using the intron in Lotus japonicus CYCLOIDEA 2 (LjCYC2) and endogenous expression of Lotus japonicus Ubiquitin (LjUbi) was examined with LjUbif/R as a control of cDNA quantity across the samples (Feng et al., 2006b). Gene expression comparisons were carried out in five species: L. japonicus and its close relative L. filicaulis (Lotus section Lotus), L. sessilifolius (section Pedrosia, closely related to the bird-pollinated group), which has a colour change with red post-change flowers, and two species from the bird-pollinated (―rhyncholotus‖) group, L. berthelotii (with red flowers) and L. maculatus (red-orange flowers). Specific primers were designed for four of the DFR copies previously isolated in L. japonicus (Shimada et al., 2005). DFR1 was amplified using the following primer pair: forward LjDFR1F (5´-GGATGAGACCTGCTGGGGTGACC-3´) and reverse LjDFR1R (5´GATTCAGGGTGCTCGAAG-3´). DFR2 using forward LjDFR2F (5´CGCCACTGTAAGAGACCCTG-3´) and reverse LjDFR2R (5´-AACATCGCTCCAGCAGCTC204  3´). DFR3 using forward LjDFR3F (5´-CTCATGGAGGGCGGCTAC-3´) and reverse LjDFR3R (5´-GATCCTTGGAATTAAAGT-3´). DFR5 using forward LjDFR5F (5´GAGAAGGTTGGTATTCAC-3´) and reverse LjDFR5R (5´-TGATGAGTGAGAGAGCAG-3´). A conserved region of the ANS gene was amplified using the following primer pair: LjANSF2 (5´-GCAGTGGGATACAATCTA-3´) and LjANSR1 (5´ATGGAGAGGTCACGCTTG-3´). This primer pair was designed based on a conserved region of the ANS from three species, L. corniculatus (AY028931) based on a partial (Paolocci et al., 2005), L. japonicus (chromosome 2, Miyakogusa.jp accession No. CM0304.350.nc) and Glycine max mRNA complete cds (EU334548). A specific primer pair was designed from a conserved region of OMT using the following primer pair: LjOMTF2 (5´-TCTGGAGACCAGTGTGTACC-3´) and LjMOTR2 (5´TGAGTCTCTTGTGGTAGTTG-3´). This primer pair was designed from a conserved region of a previously sequence from Glycine max (TIGR accession TC190220) (Kim et al., 2006), Medicago sativa CCMOT (U20736) and the best hit I found from the L. japonicus genome for this gene (chr. 4, CMO 227.500.nc + phase). Amplification of the above mentioned genes was carried out following PCR conditions: 95 C for 2 min, 30 cycles of 94 C for 45 s, 50-57 C for 1 min and 72 C 2 min, and 72 C for 2 min.  7.3 Results 7.3.1 Pigment flower composition in Lotus The flavonoids identified (cyanidin, quercetin, petunidin, peonidin and malvidin) suggest that two branches of the anthocyanin pathway, the cyanidin and delphinidin branches, are active in mature flowers at anthesis in Macaronesian Lotus (section Pedrosia), both in the yellow flowered species and in the bird-pollinated group (Table 7.3). Additionally, I identified three 205  different flavonols (or their derivatives): kaempferol, quercetin, and isorhamnetin. Based on this sampling it seems that the pelargonidin branch of the anthocyanin biosynthetic pathway is not active in mature flowers in any of the sections of Lotus tested. Furthermore, I did not observe derivatives of the delphinidin branch in the two species tested from section Lotus (L. japonicus and L. filicaulis). In my sample therefore the delphinidin pathway was restricted to section Pedrosia, whereas the cyanidin pathway was common to all species.  Pigments in species with yellow flowers but without post-anthesis flower colour modification: Yellow flowers of these species are characterized by a mixture of flavonols (mainly isorhamnetin) and traces of anthocyanin pigments. I found aglycones of delphinidin and cyanidin in three out of five species analyzed, which suggests that these branches are active even in those species that do not change flower colour after anthesis. However, the contribution of the anthocyanin pigments to overall colour in these species is likely to be minimal compared to flavonols, which are present in large amounts, and carotenoids. Additionally, the fact that anthocyanins were found in aglycone forms indicates a lack of pathway progression down these biosynthetic branches (Fig. 7.2A) (Table 7.3).  Species that modify flower colour after anthesis: Yellow flowers (pre-change) have a similar composition of flavonols (mainly isorhamnetin) to those species that do not modify flower colour, but the relative amount of delphinidin and cyanidin aglycones is relatively greater (Fig. 7.2B). In older flowers (post-change) further anthocyanidin pathway progression occurs. In addition to delphinidin and cyanidin, I found derivatives of peonidin, malvidin and petunidin in these species. The pigment composition of these post-change flowers varied accordingly with the  206  colour of each species. For instance, purple and pink flowers tend to have more derivatives of the delphinidin branch while red and orange flowers have more derivatives of the cyanidin branch. The closest related species (L. sessilifolius and L. mascaensis) of the bird-pollinated flowers have a colour change to red flowers in late anthesis. In these species, all five anthocyanins detected in the bird-pollinated species (cyanidin, delphinidin, malvidin, peonidin, petunidin) are already present in the yellow flowers but their amounts increased during flower colour modification (post-change flowers) (Fig. 7.2B and C). Flower colour change therefore appears to modify the flux into already active branches in the anthocyanin pathway; the flux of pigment production is presumably re-directed to the cyanidin and delphinidin branches instead of to flavonols, which decrease.  Red bird-pollinated flowers: these four species contain a different proportion of flavonols and anthocyanins in comparison to their relatives with yellow flowers. Bird-pollinated flowers contain mainly cyanidin derivatives and the main flavonol is quercetin (rather than isorhamnetin) (Fig. 7.2D). However, the same branches of the anthocyanin pathway are active and flavonols still contribute to flower colour in these species. However, the flux of pigment production is towards the cyanidin and delphinidin branches of the anthocyanin pathway and flavonol production is reduced, especially isorhamnetin (a derivative of quercetin). Thus, the transition from yellow beepollinated species to red bird-pollinated species in this group involves only a quantitative redirection of pigment production from flavonols to anthocyanins. These results show three major modifications of anthocyanin and flavonol composition between bee and bird-pollinated flowers: 1) a down-regulation of flavonol production, with a modification of flavonol composition, and with quercetin as the main flavonol produced in bird-pollinated flowers, 2) a down-regulation of  207  derivatives of the delphinidin branch, and 3) an up-regulation of derivatives of the cyanidin branch (mainly cyaninidin).  7.3.2 Flower reflectance and colour change perception Yellow flowers have at least three different major reflectance types (Table 7.4). We found that only seven species of this group have a reflectance type with a UV peak in the spectrum. The remaining species have a reflectance peak above 500 nm (Fig. 7.3A-C). For flowers that change colour, pre-change and post-change flowers within the same plant have a different reflectance, as expected. Flower colour modification after anthesis shifts the reflectance peak of the flowers towards the red spectrum (Table 7.4 and Fig. 7.3B and C). This change is expected to make postchange flowers less evident to bees (which cannot see red), as they will stand out less from a green foliage background (Table 7.4). We found that flowers with deep purple colours tend to have a reflectance peak above 600 nm while red-orange flowers have a peak above 500 nm. Red-orange flowers of bird-pollinated species have reflectance in the green and red receptor action spectra, with the exception of L. pyranthus which has a UV peak in early anthesis orange flowers (older flowers turn redder). Only two of the bird-pollinated species have a reflectance peak only in the red receptor action spectrum above 600 nm and are thus completely uncoloured to bees as they lack a red-light receptor (Fig. 7.3D).  7.3.3 Gene expression comparisons during flower colour modification Our expression analysis results suggest that all three structural enzymes of the anthocyanin pathway analyzed, DFR, ANS and OMT, are expressed at late developmental stages. Two copies of dihydroflavonol 4-reductase (DFR), DFR1 and DFR3, do not have differences in expression patterns among the species analyzed. DFR1 is expressed in all species while we did 208  not detect DFR3 expression in flowers of any of the species analyzed. We found differences in expression patterns in two copies of DFR. DFR2 is expressed in bee-pollinated flowers (including pre-change and post-change flowers) but it is not expressed in either of the two bird-pollinated species. The expression patterns of the four DFR copies suggest that particular copies may have specific activity on particular branches of the anthocyanin pathway. DFR5 expression was only observed in those species that produced pigments of the delphinidin branch, for which it may be specific (Fig. 7.4). Anthocyanidin synthase (ANS) is expressed in all species analyzed at approximately the same levels, even in yellow flowers (pre-change). We only observed a slight increase in expression level in red (post-change) flowers of one species (Fig. 7.4). O-methyltransferase (OMT) expression was detected in all bee-pollinated species (in both yellow pre-change and red post-change flowers). However, this enzyme seems to be downregulated in the bird-pollinated species at late developmental stages (Fig. 7.4).  7.4 Discussion 7.4.1 The changing balance of anthocyanin and flavonol pigment composition in bee and bird pollination in Lotus In many systems, transitions in pollination syndromes have involved the activation of previously absent branches of the anthocyanin biosynthetic pathway (Rausher, 2008; Scogin and Freeman, 1987; Streisfeld and Rausher, 2009b; Taylor, 1984; Zufall, 2003; Zufall and Rausher, 2003), (Smith, 2006). However, in contrast, the transition in Lotus from bee (yellow) to bird (red) flowers does not involve novel inactivation and/or activation of branches of the anthocyanin pathway. Rather, this transition involved the increasing production of anthocyanin pigments already present in small amounts, and also quantitative alterations in the proportions of flavonols, 209  such as quercetin, kaempferol and isorhamnetin. In the yellow-flowered species, isorhamnetin was often particularly abundant. Flavonols are significant in insect attraction, as they absorb UV light (Tanaka et al., 2008). Reflectance from these flowers (yellow to humans) therefore tends to stimulate the green receptor of the insect eye rather than the blue or UV receptors. Both anthocyanin and flavonol pigments are produced from the same precursor (dihydroflavonols) as branches of the same flavonoid biosynthetic pathway, and the specific production of each pigment depends on the branches of the pathway that are active (Grotewold, 2006; Tanaka et al., 2008). Most of the flux of pigment production in yellow species is towards flavonols, in particular derivatives of dihydroquercetin (DHQ), also a precursor of the cyanidin branch (Fig. 7.5). The transition in flower colour at the biochemical level therefore involves the re-direction of pigment flux production towards the cyanidin branch, which is the main anthocyanin observed in bird-pollinated flowers (Fig. 7.2d) whereas the flavonols kaempferol and isorhamnetin are greatly reduced in bird-pollinated flowers and has a minor contribution to overall pigment composition in comparison to bee-pollinated flowers (Fig. 7.2). It may be that this redirection of flux results from a small but significant shift in the competitive balance of enzymes for substrates in these branches. In lisianthus (Eustoma grandiflorum) and carnation (Dianthus caryophylus) flavonol synthase (FLS) and DFR play a critical role regulating the production of flavonols and anthocyanins, and there is a regulation to avoid competition for the substrate. In both species, the production of flavonols and anthocyanins is clearly divided, and there is no overlap in the expression of FLS with DFR (Noda et al., 2004; Stich et al., 1992; Uddin et al., 2002). In contrast to the findings for lisianthus and carnation, I did not find any clear-cut division in the production of flavonols and anthocyanins; therefore I suspect that in Lotus there is some level of competition  210  as both enzymes use the same precursor. However, the expression of FLS needs further examination in order to corroborate the findings.  7.4.2 Post-pollination floral colour change as a mechanism to promote foraging efficiency An interesting feature of Macaronesian Lotus (Lotus section Pedrosia) is the ability to change colour, possibly in response to pollination. In this particular group I found that 58% of the species have the ability to modify flower colour after anthesis. Flowers at anthesis (pre-change) are always yellow in all bee-pollinated species and post-change flowers vary from red (8 spp), orange (2 spp), pink (1 sp), brown (4 spp) to purple (6 spp) (Fig. 7.1). In all cases pre-change and post-change flowers have differences in spectral reflectance, suggesting that they are likely perceived differently by pollinating bees (Fig. 7.3B and C). This suggests that such colour change perhaps evolved as a strategy to increase foraging efficiency (Gori, 1983; Jones and Cruzan, 1982; Jones and Cruzan, 1999; Oberrath and Böhning-Gaese, 1999). Red, late anthesis, flowers may still contribute to the overall floral display to attract distant pollinators but at short range, un-pollinated flowers (with reward) can be distinguished from visited (no reward) flowers. Pollinator activity is therefore directed efficiently to nonpollinated flowers. Empirical studies of Syrmatium glabrum (=Lotus scopiarus) support the role of flower colour modification as a strategy to increase foraging efficiency (Jones and Cruzan, 1999).  211  7.4.3 Late-anthesis flower colour changes as a possible preadaptation in the evolution of the bird pollination syndrome in Lotus Red-orange bird-pollinated flowers evolved within a group of Lotus that already have a previous capacity to produce the pigments observed in bird-pollinated flowers. The five most closely related species to the bird-pollinated species have red late anthesis flowers (Fig. 7.1H). The ancestor of the bird-pollinated species was therefore likely to have had red flowers in late anthesis, a potential pre-adaptation facilitating the evolution of this pollination syndrome, as it would already have had the capacity to produce cyanidin and delphinidin derivatives. However, within Lotus as a whole the ability to change flower colour appears to be derived and has evolved at least three times within this group. Although it has been lost in some of the bird-pollinated species, which are always red/orange and do not modify flower colour after anthesis, it is retained in L. pyranthus, which starts out orange and deepens to reddish-orange (Fig. 7.6).  7.4.4 Flower colour change after anthesis as a possible facilitating factor in the evolution of bird pollination in other groups Flower colour change is not a unique feature of Lotus section Pedrosia. It is reported at least in 20% of angiosperm families (Weiss, 1995), parallel evolution that is apparently driven by the ability of colour flower change in fully turgid flowers after anthesis to act as a signal increasing efficiency in plant-pollinator interactions. Additional floral changes (such as flower orientation, size, shape and odour production) sometimes also modify in conjunction with flower colour modifications (Eisikowitch and Rotem, 1987; Raguso, 2004; Willmer et al., 2009), and it is likely that all of these modifications are an integrated system to increase foraging efficiency in some plant species (Raguso, 2008).  212  The increase of red pigmentation is one of the most common modifications in these species, perhaps because red is less visible to bees, helping to direct pollinators to non-pollinated flowers. The shift from bee to bird pollination also potentially benefits from flowers becoming less visible to bees. Therefore a heterochronic modification from red colouration in late development to red colouration early in development may be a simple mechanism to initiate change in pollination syndromes.  7.4.5 Red flowers in bird-pollinated Lotus may avoid bee visits For bees, Lotus flowers that all look the same to the human eye are surprisingly variable. Insects have three types of receptors, UV, blue and green, and they do not react to reflectance peaks above 585 nm (Chittka et al., 1994). I identified three types of reflectance pattern in yellow flowers (Table 7.4), two of which have not been reported previously in species with yellow flowers of Lotus (Floral Reflectance Database, FreD: http://reflectance.co.uk/new/) (Arnold et al., 2008). In contrast, typical red-orange bird-pollinated flowers tend to have reflectance peaks above 585 nm and are considered uncoloured to bees. These bird-pollinated flowers are not invisible to insects but are much harder to distinguish from the background green foliage (Chittka and Waser, 1997). A previous study has reported the reflectance of L. berthelotii (Ollerton et al., 2009), which is the typical reflectance of a bird-pollinated flower (Altshuler, 2003; Chittka et al., 1994). My results indicate that only two out of the four bird-pollinated species, L. berthelotii and L. eremiticus, have this pattern, with a lack of reflectance in the UV and blue, and a peak in the red range above 585 nm. The other two bird-pollinated species, L. maculatus and L. pyranthus, have significant reflectance peaks in wavelengths perceived by green visual receptors, and are similar in that regard to the bee-pollinated red flowers after late anthesis colour change (Table 213  7.4). This provides a link between the two sorts of flowers which may be significant in understanding the evolution of bird pollination in the group.  7.4.6 Down-regulation of O-methyltransferase (LjOMT) in bird-pollinated Lotus One of the main differences that we found between yellow bee-pollinated flowers and red bird-pollinated flowers is that the amount of isorhamnetin is reduced in the latter (Fig. 7.2). A flavonoid 3’ O-methyltransferase has been previously characterized in Glycine max (Kim et al., 2006). Based on this sequence, I designed primers to study the expression of the ortholog, LjOMT, in Lotus. My expression analysis of LjOMT showed a down-regulation in the two birdpollinated species that we examined (Fig. 7.4). Isorhamnetin, produced from quercetin, is the main flavonol in yellow species, and the evolutionary change in pollination syndrome involved a reduction in the production of this pigment (Fig. 7.5). These data suggest the possibility that changes in expression of this enzyme may be involved in the flux change of pigment production towards the cyanidin branch in these two bird-pollinated species (Fig. 7.5).  7.4.7 Possible specialization of dihydroflavonol 4-reductase (LjDFR) copies in Lotus Dihydroflavonol 4-reductase (DFR) is the first enzyme in the pathway specific for anthocyanin production (Grotewold, 2006) and it catalyzes the conversion of dihydroquercetin (DHQ), dihydrokaempferol (DHK) and dihydromyricetin (DHM) to uncoloured leucocyanidin, leucopelargonidin and leucodelphinidin, respectively (Fig. 7.5). This enzyme seems to be a single copy gene in many plant species examined, including arabidopsis, grape, tomato, barley, snapdragon, rice and aquilegia (Bongue-Bartelsman et al., 1994; Chen et al., 1998; Hodges and Derieg, 2009; Holton and Cornish, 1995; Kristiansen and Rohde, 1991; Sparvoli et al., 1994; Winkel-Shirley et al., 1992). Multiple copies have been found in some groups and it seems that in 214  these species only one copy is active or expressed in flowers. For instance, three copies of DFR have been isolated in Petunia hybrida, but only dfrA is expressed in flowers (Forkmann and Ruhnau, 1987). Based on substrate preferences, a region of DFR has been proposed for substrate specificity (Beld et al., 1989). An amino acid residue at position 134 of Gerbera DFR appears to determine substrate preference (Johnson et al., 2001) and two types of DFR have been identified, those having an asparagine (Asn) in this position (Asn-type DFR) and another with an aspartic acid (Asp) in this position (Asp-type DFR). Some plants such as Cymbidium (Johnson et al., 1999), Petunia (Forkmann and Ruhnau, 1987; Johnson et al., 2001) and Gentiana (Tanaka et al., 1996) have only DFRs of the Asn-type in this position and cannot produce pelargonidin-based orange-red in their flowers by converting DHK efficiently. In the legumes examined so far, between three and five copies of DFR have been found, and both types have been reported (Shimada et al., 2005; Xie et al., 2004). The maintenance of multiple copies of both types has led to the suggestion that these copies perhaps have subfunctionalized roles. In L. japonicus, it has been found that the five DFR copies have a differential expression on different tissues. For instance, the DFR3 copy is not expressed in floral tissue and is only expressed in leaves and stems (Shimada et al., 2005). In our analysis we did not find DFR3 expression in any of the five species tested (Fig. 7.4), which indicates that this copy was not likely to have been involved during the transition in flower colour. I found that DFR2 (Asn-type) is down-regulated in the two bird-pollinated species (Fig.7. 4). This copy is expressed in all three bee-pollinated species, but not across all three types of petals. In addition to flower tissue, expression of this DFR copy has been reported in pods, stems and roots of L. japonicus (Shimada et al., 2005). This enzyme has greater activity with  215  dihydrokaempferol (DHK) and dihydroquercetin (DHQ) substrates rather than with dihydromyricetin (DHM) (Shimada et al., 2005). I also found that DFR5 (Asp-type) is only expressed in those species with delphinidinbased pigments. This copy has a higher activity towards DHQ and less to DHM, and, consistent with this, I did not observe expression of this copy in L. japonicus or L. filicaulis, both species that lack delphinidin-based pigments in petals. However, a previous study reported DFR5 expression in floral tissue of L. japonicus (Shimada et al., 2005), but in this case all flower parts were included and so it is not certain that the expression was in the petals only.  216  Table 7.1 List of species in section Pedrosia (including bird-pollinated members of the rhyncholotus group) analyzed for pigment composition. Taxon  Lotus section Lotus Lotus filicaulis Durieu Lotus japonicus Gifu-B129 (Regel) K. Larsen Lotus section Pedrosia (Lowe) Christ. Lotus arinagensis Brawm.  Flower colour at anthesis  Postchanged colour  Collection information  Yellow Yellow  Red -  Cult. UBC Cult. UBC  Yellow  Red  Lotus argyrodes R.P Murray  Yellow  Purple  Lotus brunneri Webb in Hooker Lotus callis-viridis Bramwell & D.H. Davis Lotus campylocladus Webb & Berthel. Lotus dumetorum Webb ex R. P. Murray  Yellow  -  Yellow  -  Playa Arinaga, GC Cult. JCVC # 5435/UDH/07 Punta de Pargo, M Cult. JCVC # 514B/07 Sal, CV Anden Verde, GC  Yellow  -  Yellow  -  Lotus emeroides R.P. Murray Lotus glaucus Sol.  Yellow  Pink  Yellow  Orange  Lotus jacobaeus L.  Yellow  Purple  Lotus latifolius Brand  Yellow  Purple  Lotus mascaensis Burchard Lotus sessilifolius D.C. subsp. sessilifolius Lotus spartioides Webb & Berthel. Lotus tenellus (R. Lowe) Sandral, Santos & D.D. Sokoloff Lotus section Rhyncholotus (Monod) D.D. Sokoloff Lotus berthelotii Masf. Lotus eremiticus A. Santos Lotus maculatus Breitf. Lotus pyranthus P. Pérez  Yellow Yellow  Red Red  Yellow  -  Yellow  Red  Arachico San Roque, T  Red Red Red-orange Orange  Red  Cult. JAO 152-96 Cult. JAO 430-95 Cult. JAO 431-95 Cult. JAO 124/01  Voucher, herbarium  Ojeda 71/UBC Ojeda 70/UBC  Ojeda 189/UBC  Ojeda 181/UBC  Road AronaIfonche, T Road Baradero, San Andres Anaga, T Epina, G Cult. JCVC # 223/B/07 Porto Nurbita, M Cult. JCVC 166/06, Cha de Calderas, CV Cult. JCVC, Santo Antao CV Cult JCVC Guimar Poligono Industrial, T Tamadaba, GC  Ojeda 187/UBC  Marrero et al JCVC Jaén 214/03 JCVC  Ojeda 226/UBC  217  Table 7.2 Lotus species from section Pedrosia (including bird-pollinated members of the rhyncholotus group) measured for petal reflectance. Taxon Lotus section Pedrosia (Lowe) Christ. Lotus arinagensis Bramwell Lotus argyrodes R.P Murray Lotus brunneri Webb in Hooker Lotus callis-viridis Bramwell & D.H. Davis Lotus campylocladus Webb & Berthel. Lotus creticus L. Lotus dumetorum Webb ex R. P. Murray Lotus glaucus Sol. Lotus jacobaeus L. Lotus kunkelii (Esteve) Bramwell & D. H. Davis Lotus lattifolius Brand Lotus mascaensis Burchard Lotus purpureus Webb Lotus sessilifolius D.C. subsp. sessilifolius Lotus tenellus (R. Lowe) Sandral, Santos & D.D. Sokoloff Lotus section Rhyncholotus (Monod) D.D. Sokoloff Lotus berthelotii Masf. Lotus eremiticus A. Santos Lotus maculatus Breitf. Lotus pyranthus P. Pérez  Collection information  Voucher, herbarium  Cult. JCVC # 5435/UDH/07 Punta de Pargo, M Cult. JCVC # 514B/07 Sal, CV  Ojeda 189/UBC  Cult. JCVC # 64/05 Teno al to, T  Ojeda 188/UBC  Cult. JCVC # 223/B/07 Porto Nurbita, M  Ojeda 187/UBC  Cult. JCVC # 217/07  Ojeda 176/ UBC  Cult. JCVC, Santo Antao CV Punta Teno, T  Marrero et al JCVC Jaén 214/03 JCVC  Ojeda 181/UBC  Playa San Marcos, Icod de los Vinos, T Risco del Fraile, Tano, T  Cult. JAO 152-96 Cult. JAO 430-95 Cult. JAO 431-95 Cult. JAO 124/01  Ojeda 226/UBC  218  Table 7.3 Relative amounts of anthocyanidins and flavonols identified in Lotus section Pedrosia and Rhyncholotus. Relative amount was estimated based on the reflectance peak from each pigment. Pigment concentration for flowers before/after colour (in bold) change  Bird-pollinated species L. berthelotii L. maculatus L. eremiticus L. pyranthus Bee-pollinated species With flower colour change L. sessilifolius L. mascaensis L. argyrodes L. jacobaeus L. glaucus L. arinagensis L. emeroides L. tenellus Without flower colour change L. spartioides L. callis-viridis L. dumetorum L. brunneri L. campylocladus  Anthocyanins Delphinidin Malvidin  Flavonols Kaempferol  Isorhamnetin  14.63 10.68 5.24 7.76/12.40  1.93 0.10 0.14 0.34/0.62  0.70 0.55 2.65 0.13/0.32  0/0.87  2.47/3.50  0.71/0.87  6.21/14.40  0/3.99 0/39.2 0/42.5 0/1.97 0/1.62 0/3.12 0/1.75  0/2.53 0/6.98 0/40.6 0/6.58 0/3.99 0/3.97 0/3.73  4.21/4.36 1.07/1.51 2.97/15.7 12.21/23.12 6.24/14.32 7.79/8.78 13.78/11.35  2.23/2.55 2.20/24.35 2.43/5.11 1.90/4.53 0.70/4.45 1.49/1.16 1.67/0.87  15.48/15.68 3.70/15.70 4.37/5.17 6.23/15.94 2.39/8.25 5.20/33.90 6.23/2.92  0 0 0 0 0  0 0 0 0 0  7.38 2.2 3.69 3.95 3.34  2.08 0.84 1.67 2.43 2.01  9.45 1.97 8.08 1.49 5.69  Cyanidin  Peonidin  Petunidin  22.7 0.1 9.15 2.82/23.4  0.78 0 0.58 0/0.18  6.55 0.07 8.90 5.1/28.1  0.16 0 0.92 0/0  0.98 0 0.45 0/0.72  2.14/6.07  1.72/9.31  0/2.90  1.83/8.10 0/7.97 0/6.3 0/9.5 0/4.5 0/4.0 0.82/5.35  0.02/1.8 0 0/0.90 0/5.84 0/18.64 0/2.69 0/0.24 0/1.48 0/0.81  1.83/10.26 0/7.06 0/26.30 0/15.92 0/11.03 0/3.7 0.17/8.96  1.1 0.04 0.08 0 0.5  0 0 0 0 0  2.55 0.04 0 0 0.48  Quercetin  219  Table 7.4 Classification of flower reflectance of bee and bird-pollinated flowers according to human and bee perception.  Bird-pollinated species L. berthelotii L. eremiticus L. maculatus L. pyranthus Bee-pollinated species Do not change flower colour L. campylocladus L. callis-viridis L. creticus L. dumetorum L. brunneri Change flower colour L. purpureus L. latifolius L. glaucus L. kunkelii L. sessilifolius L. arinagensis L. mascaensis L. tenellus L. argyrodes L. jacobaeus  Anthesis (pre-changed) Human colour Bee-flower Reflectance perception colour  After colour change (post-changed) Human Bee-flower Reflectance colour colour perception Red Green u- b- g+ r+  Red Deep red Orange Orange  Uncoloured Uncoloured Green UV-green  u- b- g- r+ u- b- g- r+ u- b/ g+ r+ u+ b- g+ r+  Yellow Yellow Yellow Yellow Yellow  UV-green UV-green UV-green Green Green  u+ b- g+ r+ u+ b- g+ r+ u+ b- g+ r+ u- b- g+ r+ u- b- g+ r+  -  Yellow Yellow Yellow Yellow Yellow Yellow Yellow Yellow Yellow Yellow  UV-green UV-green Blue-green Blue-green Green Green Green Green Green Green  u+ b- g+ r+ u+ b- g+ r+ u- b/ g+ r+ u- b/ g+ r+ u- b- g+ r+ u- b- g+ r+ u- b- g+ r+ u- b- g+ r+ u- b- g/ r+ u- b- g+ r+  Brown Red Orange Red Red Red Red Red Purple Purple  Green Green Green Green Green Green Green Green Uncoloured Uncoloured  u- b- g+ r+ u- b- g+ r+ u- b- g/ r+ u- b- g/ r+ u- b- g/ r+ u- b- g/ r+ u- b- g/ r+ u- b- g/ r+ u- b- g- r+ u- b- g- r+  220  Figure 7.1 Red-orange flowers in the bird-pollinated species of the “rhyncholotus group”, (A) Lotus berthelotii and (B) L. pyranthus. Yellow flowers of bee-pollinated species that do not modify flower colour after anthesis (C) L. campylocladus, and late-anthetic flowers after colour change in (D) L. glaucus, (E) L. eriosolen, (F) L. jacobaeus, (G) L. emeroides and (H) L. sessilifolius.  (A)  (B)  (C)  (D)  (F)  (E)  (G)  (H)  221  Figure 7.2 Liquid chromatography-mass spectrometry (LC-MS) of bee-pollinated flowers (A) Lotus spartioides, a species with yellow flowers that do not modify flower colour after anthesis, (B) pre-change yellow flowers of L. sessilifolius, (C) post-change red flowers of L. sessilifolius and (D) a bird-pollinated species, L. berthelotii, with red flowers.  x10  6  (A)  2.5  1.0  Cyanidin  Delphinidin  1.5  Kaempferol  Quercetin  2.0  Isorhamnetin  Intens.  0.5  0.0  0  1  2  3  4  5  6  7  8  9  4  5  7  8  9  Isorhamnetin  Kaempferol  Quercetin  Peonidin  Petunidin  Malvinidin  0.75 0.50  6  Kaempferol  Quercetin  Malvinidin  1.00  3  Peonidin  (C)  1.25  2  Cyanidin  1  Cyanidin  0 0  Delphinidin  2  Delphinidin  4  Isorhamnetin  (B) 6  0.25 0.00  0  1  2  2  3  4  5  6  8  7  8  9  Isorhamnetin  1  7  Kaempferol  0.5 0  6  Malvinidin  Peonidin  1.0  0.0  5  Quercetin  1.5  Petunidin  Delphinidin  2.0  4  Cyanidin  (D)  2.5  3  9  Time [min]  222  Figure 7.3 Reflectance of yellow bee-pollinated flowers that do not change flower colour (A) Lotus callis-viridis with a UV peak (u+ b- g+ r+), a species that modifies flower colour after anthesis (B) L. latifolius (yellow and red flowers) , (C) L. sessilifolius (yellow and red flowers), and (D) reflectance of L. berthelotii, a bird-pollinated species with red flowers.  Relative reflectance  100  (A)  (B)  50  0 300  400  500  600  700  300  Wavelength (nm)  Relative reflectance  100  400  500  600  700  600  700  Wavelength (nm)  (C)  (D)  50  0 300  400  500  Wavelength (nm)  600  700  300  400  500  Wavelength (nm)  223  Figure 7.4 Gene expression comparisons of three structural genes of the anthocyanin pathway, dihydroflavonol-4-reductase (LjDFR1, 2, 3 and 5), anthocyanin synthase (LjANS) and O-methyl transferase (LjOMT) at mature stages of flower development. Ubiquitin was used as an internal control. Bee-pollinated species of L. japonicus, L. filicaulis and L. sessilifolius, the two latter with pre-change yellow flowers and post-change red flowers. Bird-pollinated species from the rhyncholotus group with red-orange flowers that do not change colour after anthesis.  Section Lotus  Stage 13 DP LP VP  Stage 13 Stage 13 DP LP VP  DP LP VP  Section Pedrosia  Stage 13  Stage 13  DP LP VP DP LP VP  Section Rhyncholotus  Stage 13 D P LP VP  Stage 13 DP LP VP  LjDFR1 LjDFR3 LjDFR2 LjDFR5 LjANS LjOMT LjUbiquitin L. japonicus  L. filicaulis  Bee-pollinated  L. sessilifolius  L. berthelotii L. maculatus  Bird-pollinated  224  Figure 7.5 Schematic representation of the anthocyanin pathway and the major modifications during the evolutionary transition from yellow flowers (bee pollination) to red flowers (bird pollination) in Lotus. Colours at the end of the major pigments indicate the colours produced for each pigment. Bold arrows indicate the pathways active in the birdpollinated species. The pelargonidin branch is inactive in the three types of petals in this group. Major transitions in pigment composition in bird-pollinated species (1) up-regulation of the cyanidin branch, (2) down regulation of the delphinidin branch and (3) down regulation of flavonol production with a sub sequent modification of flavonol composition. Gray square indicates the branch of the anthocyanin pathway with main pigment production in bird-pollinated species.  4-Coumaryl-CoA + Malonyl-Coa CHS  Isorhamnetin  Chalcone  OMT  3  CHI  Quercetin  Kaempferol  Myricetin  FLAVONOLS  Naringenin FLS DHQ Dihydroquercetin  1  F3H  F3´H  DFR (DFR1, DFR2) L-cyanidin ANS Cyanidin  Cyanidin glycoside  Peonidin glycoside  FLS  DHK Dihydrokaempferol F3´5´ H DFR L-pelargonidin ANS Pelargonidin  Pelargonidin glycoside  FLS DIHYDRO-FLAVONOLS  DHM Dihydromyricetin  2  DFR (DFR5)  L-delphinidin  LEUCOANTHO-CYANIDINS  ANS Delphinidin  ANTHOCYANIDINS  Delphinidin glycoside  ANTHOCYANINS  Petunidin glycoside  Malvidin glycoside  225  Figure 7.6 Molecular tree based on one nuclear (ITS) and three plastid regions (CYB6, trnH-psbA and matK). Character mapping of the trait flower colour change after anthesis in Lotus sections Pedrosia and Rhyncholotus. Red branches show clades where this trait has evolved and the numbers on the tree the times this trait evolved within this group (1-3). Arrows indicate the numbers of reversals, one of which occurred in three species of the rhyncholotus group. (A) represents flower colour at anthesis (pre-change) and (B) indicates flower colour after change (post-change).  226  Species that do not change flower colour Species that do change flower colour  A B  Equivocal  L. japonicus MG20 L. japonicus GIFU B129 L. burtii L. filicaulis L. corniculatus L. arborescens L. bollei L. latifolius L. purpureus L. brunneri L. jacobaeus L. eriosolen L. arenarius L. maroccanus L. arinagensis  1  L. kunkelii L. mascaensis Lotus sp. nov. 1 Tenerife L. sessilifolius L. eremiticus L. pyranthus  2  3  L. berthelotii L. maculatus L. emeroides L. argyrodes L azoricus L. macranthus L. assakensis L. erythrorhizus L. glaucus L. lancerottensis L. leptophyllus L. salvagensis L. tenellus L. callis-viridis  Bird pollination  L. genistioides L. holosericeus L. campylocladus L. hillebrandii L. spartioides L. dumetorum Anaga L. dumetorum Teno L. creticus L. pseudocreticus L. jolyi  227  7.5 Bibliography Altshuler, D.L., 2003. Flower color, hummingbird pollination, and habitat irradiance in four Neotropical forests. Biotropica 35, 344-355. Arnold, S.E., Savolainen, V., Chittka, L., 2008. FReD: The floral reflectance spectra database. Nature Proceedings. Beale, G.H., Price, J.R., Sturgess, V.C., 1941. A survey of anthocyanidins. VII. The natural selection of flower color. Proceedings of the Royal Society of London. Series B, Biological Sciences 130, 113-126. Beld, M., Martin, C.H., H., Stuitje, A.R., Gerats, A.G., 1989. Flavonoid synthase in Petunia hybrida: partial characterization of dihydroflavonol 4-reductase genes. Plant Molecular Biology 13, 491-502. Bongue-Bartelsman, M., O’Neill, S.D., Tong, Y., Yoder, J.I., 1994. Characterization of the gene encoding dihydroflavonol 4-reductase in tomato. Gene 138, 153-157. Bramwell, D., Bramwell, Z., 2001. Flores silvestres de las Islas canarias. Editorial Rueda, Madrid. Brochman, C., Rustan, O.H., Lobin, W., Kilian, N., 1997. The endemic vascular plant of the Cape Verde Islands, W Africa. Sommerfeltia 24, 213-228. Chen, M., SanMiguel, P., Bennetzen, J.L., 1998. Sequence organization and conservation in sh2/a1-homologous regions of sorghum and rice. Genetics 148, 435-443. Chittka, L., Shmida, A., Troje, N., Menzel, R., 1994. Ultraviolet as a component of flower reflections, and the color-perception of Hymenoptera. Vision Research 34, 1489-1508. Chittka, L., Waser, N.M., 1997. Why red flowers are not invisible to bees. Israel Journal of Plant Sciences. 45, 169-183.  228  Cronk, Q.C., Ojeda, I., 2008. Bird-pollinated flowers in an evolutionary and molecular context. Journal of Experimental Botany 59, 715-727. Dupont, Y.L., Hansen, D.M., Rasmussen, J.T., Olesen, J.M., 2004. Evolutionary changes in nectar sugar composition associated with switches between bird and insect pollination: the Canarian bird-flower element revisited. Functional Ecology 18, 670-676. Eisikowitch, D., Rotem, R., 1987. Flower orientation and color change in Quisqualis indica and their possible role in pollinator partitioning. Botanical Gazzete 148, 175-179. Feng, X., Zhao, Z., Tian, Z., Xu, S., Luo, Y., Ca, Z., Wang, Y., Yang, J., Wang, Z., Weng, L., Chen, J., Zheng, J., Guo, X., Luo, J., Sato, S., Tabata, S., Ma, W., Cao, X., Hu, X., Sun, C., Luo, D., 2006. Control of petal shape and flower zygomorphy in Lotus japonicus. Proceedings of the National Academy of Sciences of the United States of America 103, 4970-4975. Forkmann, G., Ruhnau, B., 1987. Distinctive substrate-specificity of dihydorflavonol 4-reductase form flowers of Petunia hybrida. . Zetschrift für Naturforschung C - a Journal of Biosciences 42, 1146-1148. Gori, D.F., 1983. Floral color change in Lupinus argenteus (Fabaceae): why should plants advertise the location of unrewarding flowers to pollinators? Evolution 43, 870-881. Grant, K.A., 1966. A hypothesis concerning the prevalence of red coloration in California hummingbird flowers. The American Naturalist 100, 85-97. Grotewold, E., 2006. The genetic and biochemistry of floral pigments. Annual Review of Plant Biology 57, 761-780. Hodges, S.A., Derieg, N.J., 2009. Adaptive radiations: from field to genomic studies. Proceedings of the National Academy of Sciences USA 106, 9947-9954.  229  Holton, T.A., Cornish, E., 1995. Genetics and biochemistry of anthocyanin biosynthesis. Plant Cell 7, 1071-1083. Jardim, R., Francisco, D., 2000. Flora endemica da Madeira. Muchia Publication. Johnson, E.T., Ryu, S., Yi, H., Oh, B.J., Cheong, H., Choi, G., 1999. Cymbidium hybrida dihydroflavonol 4-reductase does not efficiently reduce dihydrokaempferol to produce orange pelargonidin-type pigments. Plant Journal 19, 81-85. Johnson, E.T., Ryu, S., Yi, H., Shin, B., Cheong, H., Choi, G., 2001. Alteration of a single aminoacid changes the substrate specificity of dihydroflavonol 4-reductase. The Plant Journal 24, 325-333. Jones, C.E., Cruzan, M., 1982. Floral color changes in deerweed (Lotus scopiarus): possible function. Crossosoma 8, 1-6. Jones, C.E., Cruzan, M.B., 1999. Floral morphological changes and reproductive success in deer weed (Lotus scopiarus, Fabaceae). American Journal of Botany 86, 273–277. Kim, B.G., Lee, H.J., Park, Y., Lim, Y., Ahn, L.J., 2006. Characterization of an O-methyl transferase from soybean. Plant Physiology and Biochemistry 44, 236-241. Kristiansen, K.N., Rohde, W., 1991. Structure of the Hordeum vulgare gene encoding dihydroflavonol 4-reductase and molecular analysis of ant18 mutants blocked in flavonoid synthesis. Molecular and General Genetics 230, 49-59. Maddison, D.R., Maddison, W.P., 2000. MacClade 4: analysis of phylogeny and character evolution, version 4.0. Sinauer, Sunderland, Massachusetts. Maddison, W.P., Maddison, D.R., 2009. Mesquite: a modular system for evolutionary analysis. Mader, U., Podlech, D., 1989. Revision der marokkanischen Arten von Lotus L. subgenus Pedrosia (R. Lowe) Brand (Leguminosae). Mitteilungen (aus) der Botanischen Staatssammlung München 28, 513-567. 230  Monod, T., 1980. Contribution a l'étude des Lotus (Papilionaceae) ouest-sahariens et macaronésiens. Adansonia 19, 367-402. Noda, K., Glover, B.J., Linstead, P., Martin, C., 1994. Flower colour intensity depends on specialized cell shape controlled by a Myb-related transcription factor. Nature 369, 661664. Noda, N., Kanno, Y., Kato, N., Kazuma, K., Susuki, M., 2004. Regulation of gene expression involved in flavonol and anthocyanin biosynthesis during petal development in lisianthus (Eustoma grandiflorum). Physiologia Plantarum 122, 305-313. Oberrath, R., Böhning-Gaese, K., 1999. Floral color change and the attraction of insect pollinators in lungwort (Pulmonaria collina). Oecologia 121, 383-391. Olesen, J.M., 1985. The Macaronesian bird flower element and its relation to bird and bee opportunistic. Botanical Journal of the Linnean Society 91, 395-414. Ollerton, J., Cranmer, L., Stelzer, R.J., Sullivan, S., Chittka, L., 2009. Bird pollination of Canary Island endemic plants. Naturwissenschaften 96, 221-232. Paolocci, F., Bovone, T., Tosti, N., Arcioni, S., Damiani, F., 2005. Light and exogenous transcription factor qualitatively and quantitatively affect the biosynthetic pathway of condensed tannins in Lotus corniculatus leaves. Journal of Experimental Botany 56, 10931103. Proctor, M., Yeo, P., Lack, A., 1996. The natural history of pollination. Timber Press, Portland. Raguso, R.A., 2004. Why do flowers smell? The chemical ecology of fragance-driven pollination. In: Cardé, R.T., Millar, J.G. (Eds.), Advances in insect chemical ecology. Cambridge University Press, pp. 151-177. Raguso, R.A., 2008. Weak up and smell the roses: the ecology and evolution of floral scent. Annual Review of Ecology, Evolution and Systematics 39, 549-569. 231  Rausher, M.D., 2008. Evolutionary transitions in flower color. International Journal of Plant Sciences 169, 7-21. Reymaud, J., Lussignol, M., 2005. The flavonoids of Lotus corniculatus. Lotus Newsletter 35, 75-82. Sandral, G., Remizova, M.V., Sokoloff, D.D., 2006. A taxonomic survey of Lotus section Pedrosia (Leguminose, Loteae). Wulfenia 13, 97-192. Scogin, R., 1988. Floral anthocyanidins of bird-visited flowers. Botanical Gazette 149, 437-442. Scogin, R., Freeman, C.E., 1987. Floral anthocyanins of the genus Penstemon: correlations with taxonomy and pollination. Biochemical Systematics and Ecology 15, 355-360. Shimada, N., Sasaki, R., Sato, S., Kaneko, T., Tabata, S., Aoki, T., Ayabe, S., 2005. A comprehensive analysis of six dihydroflavonol 4-reductase encoded by a gene cluster of the Lotus japonicus genome. Journal of Experimental Botany 56, 2573-2585. Smith, S.D., 2006. Floral diversification and pollination biology of the Andean clade Iochrominae (Solanaceae). University of Wisconsin, Madison. Sparvoli, F., Martin, C., Scienza, A., Gavazzi, G., Tonelli, C., 1994. Cloning and molecular analysis of structural genes involved in flavonoid and stilbene biosynthesis in grape (Vitis vinifera L.) Plant Molecular Biology 24, 743-755. Stich, K., Eidenberger, T., Wurst, F., Forkmann, G., 1992. Flavonol synthase activity and the regulation of flavonol and anthocyanin biosynthesis during flower development in Dianthus caryophylus L. (carnation). Zeitschrift fuer Naturforschung, C: Biosciences 47, 553-560. Streisfeld, M., Kohn, J.R., 2005. Contrasting patterns of floral and molecular variation across a cline in Mimulus aurantiacus. Evolution 59, 2548-2559.  232  Streisfeld, M., Rausher, M.D., 2009a. Altered trans-regulatory control of gene expression in multiple anthocyanin genes contributes to adaptive flower color evolution in Mimulus aurantiacus. Molecular Biology and Evolution 26, 433-444. Streisfeld, M., Rausher, M.D., 2009b. Genetic changes contributing to the parallel evolution of red floral pigmentation among Ipomoea species. New Phytologist 183, 751-763. Suzuki, H., Sasaki, R., Ogata, Y., Nakamura, Y., Sakurai, N., Kitajima, M., Takayama, H., Kanaya, S., Aoki, K., Shibata, D., Saito, K., 2008. Metabolic profiling of flavonoids in Lotus japonicus using chromatography Fourier transform ion cyclotron resonance mass spectometry. Phytochemistry 69, 99-111. Suzuki, S., Nishihara, M., Nakatsuka, T., Misawa, N., Ogiwara, I., Yamamura, S., 2007. Flower color alteration in Lotus japonicus by modification of the carotenoid biosynthetic pathway. Plant Cell Reports 26, 951-959. Tanaka, Y., Sasaki, N., Ohmiya, A., 2008. Biosynthesis of plant pigments: anthocyanins, betalains and carotenoids. The Plant Journal 54, 733-749. Tanaka, Y., Tsuda, S., Kusumi, T., 1998. Metabolic engineering to modify flower color. Plant Cell Physiology 39, 1119-1126. Tanaka, Y., Yonekura, K., Fukuchi-Mizatuni, M., Fukui, Y., Fujiwara, H., Ashikari, T., Kusumi, T., 1996. Molecular and biochemical characterization of three anthocyanin synthetic enzymes from Gentiana triflora. Plant Cell Physiology 37, 711-716. Taylor, R.J., 1984. Floral anthocyanins of Aquilegia and their relationship to distribution and pollination biology of these species. Bulletin of the Torrey Botanical Club 111, 462-468. Uddin, A.F.M.J., Hashimoto, F., Nishimoto, S., Shimizu, K., Sakata, Y., 2002. Flower growth, coloration and petal pigmentation in four lisianthus cultivars. Journal of the Japanese Society for Horticultural Science 71, 40-77. 233  Weiss, M.R., 1995. Floral color change: a widespread functional convergence. American Journal of Botany 82, 167-185. Willmer, P., Stanley, D.A., Steijven, K., Matthews, I.M., Nuttman, C.V., 2009. Bidirectional flower color and shape changes allow a second opportunity for pollination. Current Biology 19, 919-923. Winkel-Shirley, B., Hanley, S., Goodman, H.M., 1992. Effects of ionizing radiation on a plant genome: analysis of two Arabidopsis transparent testa mutations. Plant Cell 4, 333-347. Xie, D., LJackson, L.A., Cooper, J.D., Ferreira, D., Paiva, N.L., 2004. Molecular and biochemical analysis of two cDNA clones encoding dihydroflavonol 4-reductase from Medicago trunculata. Plant Physiology 134, 979-994. Zufall, R.A., 2003. Evolution of red flowers in Ipomoea. Duke University, Durham. Zufall, R.A., Rausher, M.D., 2003. The genetic basis of a flower color polymorphism in the common morning glory (Ipomoea purpurea). Journal of Heredity 94, 442-448. Zufall, R.A., Rausher, M.D., 2004. Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428, 847-850.  234  8. Conclusion 8.1 Conclusion and future directions The study of transitions in pollination syndromes has been addressed in closely related species with contrasting flower features(see revison in Thomson and Wilson, 2008). To date several plant groups have been studied using this approach, but only a few flower features have been dissected at the genetic and molecular level (Cronk and Ojeda, 2008; Galliot et al., 2006b). The particular transition from bee (melittophily) to bird pollination (ornithophily) has been studied in at least six genera (Ipomoea, Aquilegia, Mimulus, Penstemon, Silene and Costus), all species adapted to hummingbird pollination (Thomson and Wilson, 2008). Chapter 2 is an overview of the major trends of floral features associated with bird pollination. I highlighted that some models systems are currently been used to understand the transition in pollinators (from bee to bird pollination). Of all the floral features that are modified during the this transition, flower colour is perhaps the best characterized so far (Rausher, 2008). Three genera, Ipomoea, Mimulus and Aquilegia, in particular have been used as model systems for the study of the evolution of flower colour and other features (e.g., spur length) during pollination shifts (Bradshaw and Schemske, 2003; Des Marais and Rausher, 2010; Hodges and Derieg, 2009; Rausher, 2008; Streisfeld and Rausher, 2009b). Petunia is also a promising system to study pollination transitions. To date, this system has been used to dissect the transition from diurnal bees (P. integrifolia) to nocturnal hawkmoths (P. axilaris) (Galliot et al., 2006b; Hoballah et al., 2007; Stuurman et al., 2004). Petunia exserta, a closely related species of P. axilaris, has red flowers that lack scent production and is hummingbird pollinated. This is also a particular interesting species pair currently under investigation (Hermann and Kuhlemeier, 2011). However, it is worth noting that all the systems mentioned above are hummingbirdpollinated; therefore these systems include only modification from insect to hummingbirds. Bird 235  pollination has evolved several times under the selective pressure of different types of birds (e.g., honeyeaters in Australia or sunbirds in South Africa), and therefore additional groups that have evolved similar floral features under the selection of different birds are of particular interest to further understand the evolution of ornithophily. Here, I explored the transition from melittophily to ornithophily in a group of four species of Lotus within the Canary Islands. This group of species is putatively pollinated by opportunistic passerines nectar feeders and represents a different system. This group is less studied that the other plant groups mentioned above, but it has some characteristics (easy to propagate in the nursery, short growing season and genomic resources from Lotus japonicus) that can contribute to the understanding of pollinations shifts. Most importantly, it is a system that evolved under the selection pressure other than hummingbirds, and therefore could potentially provide new insights about the evolution of bird pollination. Additional systems from other areas are needed to better understand the evolution of bird pollination. For example, it is estimated that at least 15% of the flora of southwestern Australia is bird pollinated (Keighery, 1982), and about 3-4 % (600-800 spp.) of the flora of South Africa has evolved to this syndrome (Steve Johnson, pers. comm.), yet there are no any single system under investigation under the perspective mentioned in Chapter 2. In this thesis, I particularly addressed modifications in two floral traits during the transition for bee to bird pollination: petal epidermal micromorphology and flower colour. My goal in Chapter 3 was to characterize the major epidermal types within the Leguminosae. Before my analysis, the epidermal types reported in this large family were reduced to a few species and without a phylogenetic or evolutionary context. My results provided a general overview of these epidermal types and some evolutionary trends of these epidermal types. Further analyses are needed in particular groups, especially those where I identified transitions of 236  epidermal types or the loss of characteristic epidermal types in papilionoids. More intensive sampling is also required in particular groups within this family, in particular basal papilionoids and mimosoids. Finally, modifications of epidermal types can be further analyzed in terms of transitions of pollination syndromes (e.g., in Erythrina). Chapter 4 is a combined molecular phylogeny of the Macaronesian Lotus (Pedrosia and the rhyncholotus group). My goals were to identify the closest relatives of the bird-pollinated species and to estimate the time when this syndrome evolved. I found that bird pollination evolved recently in Tenerife and La Palma, about 1.2 Ma. These four species shared a MRCA about 2 Ma with L. sessilifolius. Therefore, these results suggest that ornithophily evolved de novo in this archipelago, likely under the selection of these opportunistic passerine birds. Bird pollination has evolved in at least other five genera within Macaronesia (Olesen, 1985; Valido et al., 2004), and among these groups, it seems that bird pollination in Lotus evolved at the same period as in the Echium lineage (García-Maroto et al., 2009). The evolutionary transition of pollination syndromes in Lotus seems to be associated with the availability of niches within the last 2 Ma, in conjunction with the increase of volcanic activity in Tenerife (ca. 3 Ma) and the emergence of La Palma (1.77 Ma) and El Hierro (1.12 Ma). I also identified L. sessilifolius as the closest bee-pollinated relative of this group. Lotus sessilifolius is distributed in four islands (Hierro, La Palma, La Gomera and Tenerife) within this archipelago. However, I was unable to determine if bird pollination evolved once or twice in the rhyncholotus group or the closely related L. sessilifolius population within this species complex. More variable regions, such as ISSR, are required to fully understand the evolution of this syndrome in this group. The low levels of variability observed in this group became more evident when I tested the applicability of six gene regions (rbcL, rpoB, trnH-psbA, rpoC1, matK and ITS) as barcodes in a 237  group of 38 species. In Chapter 5, I described my findings on the applicability of these gene regions on species discrimination and conservation in this group. My goal in this chapter was only to test these regions and not to provide a set of regions that could be used in this group or in other recent island radiations. The gene regions matK + rbcL have recently been suggested for use in combination for species identification (DNA barcoding) (CBOL, 2009). The Macaronesian Lotus group represents a recent island radiation and it provides a particularly difficult challenge for barcoding. Some of these species (10) are of conservation concern and therefore barcoding has practical applications in conservation management. The intergenic region trnH-psbA was the most variable and had greatest discriminatory power (18%) of the plastid regions when analyzed alone, but was also the one that showed the most intraspecific variation, a problem that has been suggested before in other plant groups (Edwards et al., 2008). ITS was the best region of all when analyzed alone with a discriminatory power of 26% of the species discrimination. The matK region performed poorly in terms of PCR and sequencing success. The recommended combination, matK+ rbcL identifies to species level only 18% of the species and only one of the 10 endangered species. When combined in pairs, four plastid pair combinations showed slightly better discriminatory power than the recommended combination. The inclusion of ITS increased the number of species identified when combined with each of the chloroplast regions in pairs. When all regions were combined, I identified 52% of the species and 40% of the endangered or threatened species. These results indicate that novel approaches to barcoding will be needed in recently evolved groups such as those of recent island radiations. Chapter 6 explored modifications in petal epidermal surface between the two contrasting flower morphologies. Bee-pollinated species are characterized by papillose conical cells (PCS) in the exposed areas of the flower. It has been reported before that this epidermal type is widespread 238  in angiosperms (Kay et al., 1981), but a previous survey (Christensen and Hansen, 1998) was unequivocal in indicating whether bird-pollinated flowers are characterized by PCS. My goals in this chapter were to determine the petal epidermal modifications between bee- and bird-pollinated species in Macaronesian Lotus and associate this shift in pollination syndrome with the expression of three petal identity genes. I found that bird-pollinated Lotus species (rhyncholotus) have lost PCS in dorsal petals and most of the surface of the three petals is covered by tabular rugose cells (TRS). These species only have PCS in a highly localized area of the lateral petals, which is exposed to birds. The complete lack of PCS in the dorsal petal in the rhyncholotus group is associated with a down regulation of LjCYC2 (dorsal identity gene). All bee-pollinated species I analyzed have an early expression of this identity gene. My results also indicate that the closest relative of the rhyncholotus group might have a molecular pre-adaptation. LjCYC3, the lateral identity gene, is expressed early in both L. sessilifolius and the two rhyncholotus species I tested, but not in the other bee-pollinated species of Lotus. This strong early expression of the lateral petal identity gene is also associated with an increase of TRS in the rhyncholotus and its closely related species, L. sessilifolius. Therefore while a heterochronic contraction of LjCYC2 expression is directly associated with the evolution of bird pollination, a heterochronic expansion of LjCYC3 in the clade in which bird pollination evolved may have been a facilitating molecular pre-adaptation. Additional species within the L. sessilifolius group (L. arinagensis, L. mascaensis, L. emeroides and L. kunkelii) can also be analyzed to corroborate these findings. Besides this study, there are no other similar analyses that compared closely related species with contrasting morphologies; therefore more examples are needed to determine if transitions from bee to bird pollination are characterized by a lack of PCS. The five species of Erythrina I analyzed in Chapter 3 have papillose conical cells (either PKR or PCS) on all three types of petals. However, Erythrina is 239  completely pollinated by birds, either hummingbirds or passerine birds (Bruneau, 1997), therefore more detailed analyses from sister species with contrasting pollination syndromes are further required. Species pairs within Ipomoea, Aquilegia, Penstemon, Mimulus, Silene, Salvia and Costus where several transitions have occurred within each genus are of potential interest. Additionally, a more widespread analysis, including species from several regions can also provide a more complete picture of the trends in petal epidermal micromorphology modifications during pollinator transitions. Additional genes known to be involved in PCS differentiation (such as MIXTA in snapdragon) (Comba et al., 2000) and trichome differentiation (e.g., GLABRA1) deserve further examination within this group. Homologues copies of MIXTA have been isolated in Legumes (Beverly Glover, pers. comm.) and could provide a better understanding of the transition of epidermal types in Macaronesian Lotus. . Finally, in Chapter 7 I studied modifications in flower color between the two syndromes. Red flowers are one of the main floral traits usually associated with bird pollination. In Macaronesian Lotus, bird-pollinated species are red-orange while bee-pollinated are yellow, at least before modifying flower color after anthesis. I found that 58% of the species within Pedrosia have the ability to modify flower color after anthesis. Yellow flowers can change to red, purple, brown, orange and pink post-anthesis colours, likely as a strategy to increase foraging efficiency of insect pollinators, as shown in other plant groups (Eisikowitch and Rotem, 1987; Gori, 1983; Jones and Cruzan, 1982; Oberrath and Böhning-Gaese, 1999; Willmer et al., 2009). My objectives in this chapter were: (1) to determine if the modification in flower colour is perceived differently by pollinators, (2) to establish the pigments modifications during the 240  transition to bird pollination, and (3) to determine whether genes in the biosynthetic pathway have differential expression patterns during the transition. My reflectance analyses from yellow and post-anthesis flower colors within the same species showed that the two colour types are perceived differently by bees, and likely this affects flower visitation. The red-orange bird-pollinated flowers have the typical reflectance reported for other bird-pollinated species, which suggests that bees will have difficulty to distinguish them from the green background (Arnold et al., 2008; Arnold et al., 2010; Chittka et al., 1994; Chittka and Waser, 1997; Ollerton et al., 2009). This evidence suggests that the red/orange colours of these bird-pollinated flowers may have evolved as a deterrent (anti-bee) trait, rather to attract birds. From my results of the mapping of the evolution of flower colour, it seems that the clade in which bird pollination evolved had a pre-adaptation to modify flower color. I found that the pigments involved in the modification of these colours are already present in the yellow flowers before the modification of flower colour. The modification of flower colour within the beepollinated species is due to an increase of anthocyanidin production with no modification in flavonol composition. In contrast, the transition from bee to bird pollination required the redirection of anthocyanidin production (especially the cyanidin branch) together with a modification of flavonol composition (the main flavonol in bee-pollinated species is isorhamnetin); birdpollinated flowers have more quercetin than isorhamnetin derivatives. Therefore, the rhyncholotus group evolved the red /orange flowers maintaining the same pathways active but modifying the flux and composition between anthocyanidins and flavonols. Thus, the evolution of red/orange flowers in bird-pollinated species appears to be a heterochronic modification of  241  pigment production. Bird-pollinated species have red flowers from early developmental stages, instead of a late expression due to modification of flower color after anthesis. On the other hand, transitions from blue (bee-pollinated) to red (bird-pollinated) in species of Ipomoea, Mimulus and Penstemon (Rausher, 2008) required the inactivation/activation of different branches of the anthocyanidin pathway. In Ipomoea, for instance, the cyanidin branch (responsible for blue pigments) is inactivated and the down regulated f3h gene has apparently accumulated structural mutation, while DFR has evolved a greater specificity to the precursor leading to pelargonidin (responsible for red colors) (Zufall and Rausher, 2004). In the particular case of Macaronesian Lotus, the expression patterns of ANS and DFR are in agreement with the pigments I identified in the petals, providing further evidence that in this particular group only two branches, the cyanidin and delphinidin are active in this group. I also found evidence that might suggest that LjOMT, responsible to convert quercetin to isorhamnetin, is down regulated in the two bird-pollinated species. Finally, it seems that there is a specialization in some of the different copies of DFR in Lotus. I found that DFR2 is down regulated in the two bird-pollinated species and might indicate that this copy has more specificity to the delphinidin branch. The expression patterns reported here should be considered as preliminary results and further analyses are required to fully understand this colour modification. One future avenue of further research could be the inclusion of FLS, responsible for flavonol production, as well as the further characterization of ANS in Lotus. Only DFR has been fully characterized in Lotus japonicus (Shimada et al., 2005), however, it is not known yet the number of copies in LjOMT and ANS..  242  8.2 Bibliography Arnold, S.E., Savolainen, V., Chittka, L., 2008. FReD: The floral reflectance spectra database. Nature Proceedings. Arnold, S.E.J., Faruq, S., Savolainen, V., McOwan, P., Chittka, L., 2010. FReD: the floral reflectance database--a web portal for analyses of flower colour. PloS one 5, e14287. Bradshaw, H.D.J., Schemske, D.W., 2003. Allele substitution at a flower color locus produces a pollinator shift in monkeyflowers. Nature 426, 176-178. Bruneau, A., 1997. Evolution and homology of bird pollination syndromes in Erythrina (Leguminosae). American Journal of Botany 84, 54-71. CBOL, 2009. A DNA barcode for land plants. Proceedings of the National Academy of Sciences of the USA 106, 12794-12797. Chittka, L., Shmida, A., Troje, N., Menzel, R., 1994. Ultraviolet as a component of flower reflections, and the color-perception of Hymenoptera. Vision Research 34, 1489-1508. Chittka, L., Waser, N.M., 1997. Why red flowers are not invisible to bees. Israel Journal of Plant Sciences. 45, 169-183. Christensen, K., Hansen, H., 1998. SEM-studies of epidermal patterns of petals in the angiosperms. Opera Botanica 135, 1-91. Comba, L., Corbet, S., Hunt, H., Outram, S., Parker, J.S., Glover, B.J., 2000. The roles of genes influencing the corolla in pollination of Antirrhinum majus. Plant, Cell and Environment 23, 639-647. Cronk, Q.C., Ojeda, I., 2008. Bird-pollinated flowers in an evolutionary and molecular context. Journal of Experimental Botany 59, 715-727. Des Marais, D.L., Rausher, M.D., 2010. Parallel evolution at multiple levels in the origin of hummingbird pollinated flowers in Ipomoea. Evolution 64, 2044-2054. 243  Edwards, D., Horn, A., Taylor, D., Savolainen, V., Hawkins, J.A., 2008. DNA barcoding of a large genus, Aspalathus L. (Fabaceae). Taxon 57, 1317-1327. Eisikowitch, D., Rotem, R., 1987. Flower orientation and color change in Quisqualis indica and their possible role in pollinator partitioning. Botanical Gazzete 148, 175-179. Galliot, C., Stuurman, J., Kuhlemeier, C., 2006. The genetic dissection of floral pollination syndromes. Current Opinion in Plant Biology 9, 78-82. García-Maroto, F., Mañas-Fernández, A., Garrido-Cárdenas, J.A., Alonso, D.L., Guil-Guerrero, J.L., Guzmán, B., Vargas, P., 2009. Delta6-desaturase sequence evidence for explosive Pliocene radiations within the adaptive radiation of Macaronesian Echium (Boraginaceae). Molecular Phylogenetics and Evolution 52, 563-574. Gori, D.F., 1983. Floral color change in Lupinus argenteus (Fabaceae): why should plants advertise the location of unrewarding flowers to pollinators? Evolution 43, 870-881. Hermann, K., Kuhlemeier, C., 2011. The genetic architecture of natural variation in flower morphology. Current Opinion in Plant Biology 14, 60-65. Hoballah, M.E., Gubitz, T., Stuurman, J., Broger, L., Barone, M., Mandel, T., Dell'Olivo, A., Arnold, M., Kuhlemeier, C., 2007. Single gene-mediated shift in pollinator attraction in Petunia. Plant Cell 19, 779-790. Hodges, S.A., Derieg, N.J., 2009. Adaptive radiations: from field to genomic studies. Proceedings of the National Academy of Sciences USA 106, 9947-9954. Jones, C.E., Cruzan, M., 1982. Floral color changes in deerweed (Lotus scopiarus): possible function. Crossosoma 8, 1-6. Kay, Q.O.N., Daoud, H.S., Stirton, C.H., 1981. Pigment distribution, light reflection and cell structure in petals. Botanical Journal of the Linnean Society 83, 57-84.  244  Keighery, G.J., 1982. Bird-pollinated plants in Western Australia. In: Amnstrong, J.A., Powell, J.M., Richards, A.J. (Eds.), Pollination and evolution. Royal Botanical Garden, Sidney. Oberrath, R., Böhning-Gaese, K., 1999. Floral color change and the attraction of insect pollinators in lungwort (Pulmonaria collina). Oecologia 121, 383-391. Olesen, J.M., 1985. The Macaronesian bird flower element and its relation to bird and bee opportunistic. Botanical Journal of the Linnean Society 91, 395-414. Ollerton, J., Cranmer, L., Stelzer, R.J., Sullivan, S., Chittka, L., 2009. Bird pollination of Canary Island endemic plants. Naturwissenschaften 96, 221-232. Rausher, M.D., 2008. Evolutionary transitions in flower color. International Journal of Plant Sciences 169, 7-21. Shimada, N., Sasaki, R., Sato, S., Kaneko, T., Tabata, S., Aoki, T., Ayabe, S., 2005. A comprehensive analysis of six dihydroflavonol 4-reductase encoded by a gene cluster of the Lotus japonicus genome. Journal of Experimental Botany 56, 2573-2585. Streisfeld, M., Rausher, M.D., 2009. Genetic changes contributing to the parallel evolution of red floral pigmentation among Ipomoea species. New Phytologist 183, 751-763. Stuurman, J., Hoballah, M.E., Borges, L., Moore, J., Basten, C., Kuhlemeier, C., 2004. Dissection of floral pollination syndrome in Petunia. Genetics 168, 1585-1599. Thomson, J.D., Wilson, P., 2008. Explaining evolutionary shifts between bee and hummingbird pollination: convergence, divergence, and directionality. International Journal of Plant Sciences 169, 23-38. Valido, A., Dupont, Y.L., Olesen, J.M., 2004. Bird-flower interactions in the Macaronesian islands. Journal of Biogeography 31, 1945-1953.  245  Willmer, P., Stanley, D.A., Steijven, K., Matthews, I.M., Nuttman, C.V., 2009. Bidirectional flower color and shape changes allow a second opportunity for pollination. Current Biology 19, 919-923. Zufall, R.A., Rausher, M.D., 2004. Genetic changes associated with floral adaptation restrict future evolutionary potential. Nature 428, 847-850.  246  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0071648/manifest

Comment

Related Items