sign in sign up
<<
Home , Alangium, Anagallis, Ardisiandra, Ericales, Gironniera, Illiciaceae

            

     

    !  " #$ !%

 !

         "#"$"    % &'(&"##'&'  ) *+**))* ",,'$                   !" # $ %&''( ')''*    *   * +  ,-  .    ,

 /0.1,&''(,  2 $     +  :2   "   3"0-"4 " 3 4,"   ,               567,6! , ,(8!1( 1%%7187(816,

-** * .      * *"0-"    *      3  0-   " 4  0  2 , 2    9   .  . :    / :     * , #        0.1 ,0 2   *   .               , +            .                    *             ,#  *    * " .       *  ".* *       "  , 0    * *         *       3    ! "#    4.   ,-            ;.    *  .        "3-20-24   "3 5 &'1 &  5 ;1" 21<4=  ,-       , ,    *  ,2 *      .      ,"*    * , ,     #    ;.        , $        1              *    * ,-                   "     "     .      . ;.    ,;     *        ,

$%"0-"    "-2*  ;.  00-2Lysimachia$  "   rpl 5rpl&'1rps &rps 5 2      trn;1psb"trn21<

& ' (Department of    )  ( *  )   Centre (EBC)( Norbyvägen 18D, Uppsala University, SE-75236 Uppsala, Sweden and Department of Phanerogamic Botany, Museum of Natural History P.O.( Box 50007, $ >01 /&''(

022 5% 15& 7 02(8!1( 1%%7187(816  )  ))) 1 ''68&3 )?? ,@,? A B )  ))) 1 ''68&4 List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals:

I Oh, I.-C., Denk, T., and Friis, E. M. 2003. Evolution of (): Mapping morphological characters on the molecular . Systematics and Evolution 240: 175-209.

II Denk, T. and Oh, I.-C. 2005. Phylogeny of based on morphological data: evidence from modern and the fos- sil record. Plant Systematics and Evolution 256: 113-145.

III Oh, I.-C., Anderberg, A.-L., Schönenberger, J. and Anderberg, A. A. 2008 Comparative morphology and character evolution in the Lysimachia (Myrsinaceae) and related taxa. Plant Systematics and Evolution 271: 177-197.

IV Oh, I.-C., Schönenberger, J., Motley, T. J., Myrenås, M. and An- derberg A. A. Phylogenetic relationships among endemic Hawai- ian Lysimachia (Myrsinaceae): insights from nuclear and chloro- plast DNA sequence data. Manuscript.

Papers I–III have been reproduced with permission from the publisher.

In the papers included in this thesis Il-Chan Oh has been responsible for the SEM studies of , molecular lab work and the writing, with comments and suggestions given by the co- authors.

Cover photos: Seeds of ANITA members and Lysimachia (seed) Lysimachia mauritiana (surface) (seed) Illicium floridanum (hilar area) glabra (seed) Schisandra glabra (hilar area)

Contents

Introduction...... 7 ...... 9 Lysimachia ...... 9 Aims...... 11 Summary of papers ...... 12 Paper I ...... 12 Paper II ...... 14 Evidence from the remaining ANITA members ...... 16 Discussion of papers I and II...... 22 Paper III...... 24 Paper IV ...... 27 Sammanfattning (Swedish summary) ...... 29 Acknowledgements...... 32 References...... 34

Abbreviations

cpDNA chloroplast DNA ETS external transcribed spacer ITS internal transcribed spacer nrDNA nuclear ribosomal DNA rpl16 ribosomal protein L16 rpl20-rps12 ribosomal protein L20-ribosomal protein S12 rps16 ribosomal protein S16 SEM scanning electron microscope

Introduction

The seed is one of the most remarkable innovations in the evolution of vas- cular plants. It is a significant unit for reproduction and dispersal providing protection to the embryo and supplying the growing embryo with nutrients. Seeds can tolerate harsh environments such as cold and dry conditions and survive during a dormant period in the soil for many years before germinat- ing (Ingrouille 1992). Within the seed plants two major groups are recog- nized, gymnosperms with naked and angiosperms with ovules en- closed in a carpellary tissue. In gymnosperms the embryo is nourished by the mother plant during its development and a nutritious tissue (secondary en- dosperm) surrounding the embryo is a further innovation in angiosperms that supports germination and early growth of the seedling and may be one of the factors that promoted the evolutionary success of the angiosperm. Today angiosperms are the most successful and highly diversified plant group com- prising about 225.000-350.000 extant (Kubitzki 1993). The recognition of evolutionary signals using only seed characters is in- tricate because there are no congruent evolutionary trends indicating which type of seed characters are primitive or derived, and some morphological characters are hard to polarize without detailed ontogenetic investigation. However, seed morphology has previously been used as a source of informa- tion about the evolution of angiosperms (Bouman 1974, Corner 1976, Neto- litzky 1926, Takhtajan 1988, Williams and Friedman 2002). Ovules and seeds can be characterized in terms of their organization (e.g., orthotropous, anatropous, campylotropous), number of integuments (uniteg- mic, bitegmic), structure of seed wall (e.g., testal, exotestal, endotestal, teg- mic) and a variety of morphological characters such as shape, size and sur- face pattern of seeds, hilum characters, micropyle, chalaza, raphe, and vari- ous subsidiary appendages (Fukuhara 1999, Werker 1997). The /seeds may be uncurved with the micropyle at one end and the attachment at the other (orthotropous, Fig. 1A). This type is rare in angio- sperms and mainly found in one-seeded , such as in , Piperaceae, Juglandaceae, Urticaceae, and Polygonaceae (Boesewinkel and Bouman 1984, Endress 1987, Endress 1994). Among the members of the ANITA grade (, , , , and Aus- trobaileyaceae) Amborella has almost orthotropous ovules (Endress 1994, Endress and Igersheim 1997, Endress and Igersheim 2000). From the ortho- tropous type various more complicated types can be derived, but this does

7 not necessarily mean that the orthotropous type is the ancestral one within the angiosperms (cf. Bocquet and Bersier 1960 versus Boesewinkel and Bouman 1984). The most common ovule type within the angiosperms is the anatropous type (about 80%; Boesewinkel and Bouman 1984), where the ovule is curved with the micropyle facing the placenta close to the funicle (Fig. 1B). This type is realized in the vast majority of angiosperms and oc- curs across the different orders of flowering plants. Among anatro- pous ovules are the rule in, for example, , in families such as Myrsi- naceae, , and Theophrastaceae (Johri et al. 1992). Among the ANITA members (including ), Illiciaceae, Trimeniaceae, and Austrobaileyaceae have anatropous ovules and seeds. Schisandraceae also have anatropous ovules (Igersheim and Endress 1997), but seeds are campylotropous. In campylotropous ovules the ovule is also curved, but the funicle is attached midway between the chalaza and micro- pyle (Fig. 1C). This appears to be a rather derived seed type. It is found also in Papaveraceae, Capparaceae, Cactaceae, Fabaceae, Geraniaceae, and Mal- vaceae among others (Boesewinkel and Bouman 1984). In some cases, as in Schisandraceae, campylotropous seeds may develop from anatropous ovules.

Figure 1. Different ovule and seed types found in ANITA members. A. Orthotro- pous ovule. B. Anatropous ovule. C. Campylotropous ovule. D. Longitudinal of Illicium ovule. E. Longitudinal section of Schisandra seed. A, B, C redrawn from Bocquet and Bersier (1960), D taken from Buxbaum (1961), E taken from Corner (1976).

8 Basal angiosperms The recognition of a basalmost group of extant angiosperms has been one of the main issues in plant biology for more than a century (Arber and Parkin 1907, Cronquist 1981, Takhtajan 1997). Phylogenetically, basal angiosperms comprise more than 30 families including Amborellaceae, - ceae, Chloranthaceae, Illiciaceae, Lauraceae, Magnoliaceae, Nymphaeaceae, Piperaceae, Schisandraceae, Winteraceae and so on. Several different groups have been suggested as sister to the rest of the an- giosperms based on molecular phylogenetic analyses as well as morphological studies. were regarded as a basal clade among angiosperms based on morphological analyses (Donoghue and Doyle 1989). Analysis of cpDNA rbcL sequences recognized the Ceratophyllum (Ceratophylla- ceae) as the sister to all other extant angiosperms (Chase et al. 1993, Qiu et al. 1993). Nymphaeaceae, monocots, , and Aristolochiaceae were recog- nized as successive sister groups by combined analyses using both morpho- logical and rRNA sequence data (Doyle 1994). Recent molecular and morpho- logical phylogenetic studies suggested the so-called ANITA grade comprising Amborellaceae, Nymphaeaceae (including Cabombaceae), and Illiciaceae- Schisandraceae-Trimeniaceae-Austrobaileyaceae as the basalmost lineages within the angiosperms. The genus Amborella has been suggested as sister to all other extant angiosperms (Qiu et al. 1999, Qiu et al. 2000, Soltis et al. 2000, Soltis et al. 1999, Zanis et al. 2002, Zanis et al. 2003). While some of the ANITA members are monotypic or very species poor and restricted to small geographical areas at present (Amborella, ; Austrobai- leya, Queensland; , Sulawesi, , southwestern Pacific), others display a typical East Asian-North American disjunction and are moderately specious (Illiciaceae, Schisandraceae), and Nymphaeaceae are cosmopolitan consisting of about 60 to 70 herbaceous species. Very recently the aquatic Hydatellaceae, previously thought to be a and with the sole genus Trithuria comprising a dozen spe- cies in Australasia and India, has been shown to be a member of Nym- phaeales (Friis and Crane 2007, Saarela et al. 2007, Sokoloff et al. 2008). The ovules in Hydatellaceae are anatropous, as in most other members of ANITA.

Lysimachia Comprising more than 150 species, Lysimachia L. is a genus of perennial, biennial or annual herbs, or sometimes (Hawaiian spp.) with an almost worldwide distribution (Ståhl and Anderberg 2004), but with few species in and its main center of diversity in south-western , where about two-thirds of the species are found (Chen and Hu 1979, Hu and Kelso 1996).

9 Lysimachia has traditionally been regarded as a member of the family Primulaceae. However, based on phylogenetic analyses using molecular as well as morphological data Lysimachia, along with several other genera, is now often placed in Myrsinaceae (Anderberg and Ståhl 1995, Anderberg et al. 1998, 2002, Källersjö et al. 2000, Ståhl and Anderberg 2004). The mono- phyly of the genus and its infrageneric relationships have been studied by Hao et al (2004), Manns and Anderberg (2005), and Andeberg et al (2007). One of the monophyletic groups within Lysimachia that is still not well understood from a phylogenetic point of view is subgenus Palladia, a group of species with white or sometimes pink or red arranged in terminal racemes, and characteristic reticulate seeds. A close relationship between the species of subgenus Palladia and the endemic Hawaiian species of subgenus Lysimachiopsis has been suggested based on morphological and molecular data (Marr and Bohm 1997, Hao et al. 2004, Anderberg et al. 2007). How- ever, it is still unclear whether they are sister groups or if one is part of the other (Anderberg et al. 2007).

Figure 2. , a member of L. subgen. Palladia. Photo by Arne Anderberg.

10 Aims

In papers I and II, I did comparative seed morphological studies in selected members of the ANITA grade. I wanted to know how character states are distributed within different members of ANITA, and whether or not seed characteristics in ANITA members are markedly ancestral. Using a (molecu- lar) phylogenetic framework, I mapped morphological characters on phy- logenetic in order to establish possible evolutionary or adaptive trends within the two closely related families Illiciaceae and Schisandraceae. In p. 16–21 in the thesis information is also provided on seed morphology of other members of ANITA. I used the literature and re-studied some to see whether ANITA type seeds occur in the record, mainly in the earliest records of angiosperms. Based on the results of my studies I discussed the usability of single morphological character complexes to infer plant phylog- eny. Paper III focuses on Lysimachia and closely related taxa. The aims are (1) to describe and document variation in seed morphology in a group of higher angiosperms, (2) to identify possible synapomorphies that are con- gruent with well-supported clades in the Lysimachia phylogeny, and (3) to reconstruct the evolution of seed characters. Paper IV presents phylogenetic relationships among the endemic Hawai- ian Lysimachia species using nrDNA (ETS, ITS) and cpDNA (rpl16, rpl20- rps12, rps16, trnH-psbA, trnS-G) sequence data. The objectives of paper IV are (1) to elucidate phylogenetic relationships among the endemic Hawaiian Lysimachia species (subgenus Lysimachiopsis), (2) to compare and discuss our results with respect to previous classifications, biogeography, and mor- phology, particularly with respect to the L. hillebrandii/remyii complex, and (3) to provide a phylogenetic basis for an improved of subgenus Lysimachiopsis.

11 Summary of papers

Paper I In my first paper, I studied how seed morphological characters reflect rela- tionships in Illiciaceae and how morphological characters are distributed when they are mapped on a molecular tree (Fig. 3). I also did a re-evaluation of the fossil record, i.e. seeds and fruits which previously have been ascribed to Illiciaceae, and established their affinities to modern species/species groups of Illicium. Illiciaceae consists of only one genus, Illicium, with about 35 species, which are distributed in East with about 30 species and southeastern , Mexico, and the West Indies with about 5 species (Keng 1993, Lin 1997, Smith 1947). Traditionally, Illiciaceae were included in Illiciales together with Schisandraceae and they have even been classified in Magnoliaceae (Rehder 1940). According to the updated APG classification, Illiciaceae is a member of (sensu APG II: Angiosperm Phy- logeny Group, 2003) together with Austrobaileyaceae, Schisandraceae, and Trimeniaceae (APG 2003). Investigation of morphological characters in basal angiosperms is particu- larly interesting and can provide useful data for a better understanding of phylogeny and character evolution in flowering plants, because they can be optimized on phylogenetic trees obtained from molecular studies. Recently, a nuclear rDNA ITS analysis of 15 species of Illcium suggested a major split between modern members of Illicium into an East Asian and North Ameri- can clade (Fig. 1 in Hao et al. 2000). Investigation of seed morphological characters in Illiciaceae in a molecu- lar phylogenetic framework showed some significant characters in the struc- ture of the hilar area. Most conspicuous is the hilar rim only found in North American species of Illicium, which would support a major division between East Asian and North American species. Mapping morphological characters on the molecular tree showed that three synapomorphies from seed morphological characters support the North American clade but several other features are homoplastic. In addition, a number of fossil seeds previously described as Illiciaceae were re-evaluated. Illicium germanicum, formerly described as very closely related to modern Illicium, differs from Illicium in many characters and should be excluded from this genus. Moreover, we discussed some seed

12 characters of modern and fossil species of Illiciaceae, morphological rela- tionship within the ANITA taxa, and the modern disjunction of Illicium in a biogeographical context.

61112 13 4 5 7 8 10 14 16 17 I. floridanum 99 113 2 2611 13 1 1 1 1 1 0 1 1 1 I. parviflorum 0 0 1 0 8 I. anisatum 73 1 77 14 3 6 10 12 15 I. simonsii 01 2a 1 0 0 1 1 14 17 I. angustisepalum 1 1 1 3 9 15 19 65 55 I.dunnianum 1 0 0 0 1 0 2 910111314

11121201 0 I. dunnianum 2 57 18 19 I. spathulatum 2b 1 1 10 11 13 14 I. arborescens 99 1 12 2 1 4 5 6 7 8 101113141617 3 10 11 13 I. verum 0 0 2 0 0 1 2 1 0 00 1 2 0 1 28910111617 I. majus 14 83 2 112 121 1 13 11 55 01 I. lanceolatum 0 1 6 10 I. henryi 3 59 1 0 13 I. micranthum 01

I. difengpi

Kadsura Figure 3. Morphological characters optimized on the molecular ITS tree of Hao et al. (2000): 1, 2a, 2b, and 3 correspond to clades referred to in the text. Black circles indicate apomorphies with a consistency index 1.0 and white circles characters with homoplasy (consistency index below 1.0). Bootstrap values from 100 replicates are indicated as bold numbers above branches.

13 Paper II

In the second paper I studied seed morphology of around 20 species of Schi- sandraceae. At the same time epidermal characters were studied mainly by the first author. The aim of this paper was to see how adding seed and leaf epidermal characters to a recently published data matrix of morphologi- cal characters would effect the phylogeny, resolution, and support of the phylogenetic trees obtained from the first analysis. Mainly, I was interested in whether or not additional morphological data would corroborate mono- phyly of the two currently recognized genera Kadsura and Schisandra with- in the Schisandraceae. Basically, the extremely elongated receptacle with widely spaced carpels is characteristic of Schisandra, whereas the receptacle in Kadsura is short with densely spaced carpels. Furthermore, the carpel in Schisandra is conspicuously constricted at the point of its attachment to the axis, whereas a pseudostigma occurs only in Kadsura. These features sug- gest the two genera to be monophyletic. Seed morphological data added to Hao et al.’s (2001) matrix did not provide further evidence (i.e. stronger support) for monophyly of the two genera. Instead, a number of characters appear to support the relationship of pairs of species, e.g. K. coccinea and K. scandens, both of which have conspicuously large seeds along with a multi- layered mesotesta. Among leaf epidermal characteristics, species of Kadsura were found to be consistently amphistomatic, whereas species of Schisandra have stomata only on the abaxial leaf surface. In conclusion, Kadsura and Schisandra appear to be supported as monophyletic sister taxa by a number of synapomorphies in reproductive and vegetative organs (Fig. 4). Fossils ascribed to Schisandraceae are rare in the Cretaceous. While leaf remains are difficult to ascribe to Schisandraceae because of the lack of syn- apomorphies for the family, appear to prove the presence of the fam- ily in the Late Cretaceous. In the Early Cainozoic, leaf and seed remains from North America and unambiguously belong to the family. Seeds from the of North America show some similarities to the modern Schisandra glabra from North America. By contrast, fossils from Europe show more similarities to modern Asian species.

14 I. angustisepalum

I. dunnianum

75 18 26 2740 34 20 22 23 25 28 29 30 K. coccinea A 011 1 3 0 13 1 2 1 2 2 38 5 K. scandens B 64 01 1

14 18 27 35 38 31 36 4 K. longipedunculata 30 11210 11 1 74 01 K. oblongifolia 11 20 34 40 111 K. angustifolia C 01 24 38 55 3222330 21 99 11 K. heteroclita 37 112011 1 12 15 22 23 242934 1 25 36 1210012 K. japonica 01 32 S. grandiflora 1213 15 28 91 1 D 22 29 30 38 41 0111 S. rubriflora 12 12 1 12 01 58 26 3137 39 16 17 19 20 27 S. chinensis E 8 10 10 11 2

1121 10 25 30 32 3338 6 9 16 25 S. glabra F 1011 1 2 1 1 52 00 3 27293440 8 G 14 31 S. propinqua 11212141 2 22 3138 39 2 1 S. pubescens 12 0 0 1 741 29 37 S. glaucescens 11 12 1 30 29 38 H S. sphenanthera 1 12 12 26 38 41 69 30 37 120 S. arisanensis 28 33 01 22 29 41 1 2 11 S. henryi 12 12 01 1 1 Figure 4. One of 34 most parsimonious trees recovered from a branch-and-bound search. Morphological characters are mapped on the tree using the resolving option ACCTRAN. Black circles indicate apomorphies with a consistency index 1.0 and white circles characters with homoplasy (consistency index below 1.0). Autapomor- phies are indicated following the species name. Numbers above circles refer to the character numbers, below circles to character states. Bootstrap support (%) from 1000 replicates is indicated above branches. Tree length is 107, CI = 0.664, HI = 0.561, RI = 0.700, RC = 0.464. Branches marked with a star collapse in the strict consensus tree. A = K. subgen. Cosbaea, B = K. subgen. Sarcocarpum, C = K. subgen. Kadsura, D = S. subgen. Pleiostema, E = S. subgen. Schisandra sect. Maximowiczia, F = S. subgen. Schisandra sect. Schisandra, G=S. subgen. Schisandra sect. Sphaerostema, H=S. subgen. Sinoschisandra (classification follows Saunders 1998, 2000)

15 Evidence from the remaining ANITA members 1. Amborellaceae (Pl. 1, Figs. 1-4) Specimens examined: Amborella trichopoda Baill.; J. T. Johansson 62 (S), H. S. MacKee 5181 (K) The gynoecia of Amborellaceae are apocarpous composed of 5-8 free, spi- rally arranged carpels (Endress 2001, Endress and Igersheim 2000). Each carpel contains a single, almost orthotropous (Pl. 1, Fig. 2), bitegmic and crassinucellate ovule (Endress 2001). The micropyle is formed by the lobed, annular inner integument (Endress and Igersheim 1997). The outer integu- ment is annular and sometimes lobed (Endress and Igersheim 1997). The fruitlets are drupes, 6-9 mm long (including 0.5-1 mm stalk) and 3.5-4 mm broad. The epicarp of the drupe has a more or less smooth surface (Endress and Igersheim 2000) and the endocarp is sclerotic with a reticulate-rugose surface (Pl. 1, Fig. 1). Seeds one per carpel, ovoid in shape with a beak-like remnant of the funi- cle (Pl. 1, Fig. 3), close to dark brown in colour. Dimensions of seed: length 3.5-3.6 mm; width (dorsiventrally) 1.4-1.6 mm; thickness (lateral width) 1.0- 1.2 mm. Outer surface of testa is more or less rugose under LM and reticu- late near micropylar area under SEM (Pl. 1, Fig. 4). The outer surface in the middle part of seed and funiclar area is composed of many cells, which have space between cells. Testa is about 10 m thick and membranous. It is 3-4 cell layers thick in the ovule (Endress and Igersheim 1997). Tegmen is two or three cell layers in the ovules (Endress and Igersheim 1997). In mature seeds the tegmen has two cell layers. Micropylar region is about 70 m in diameter (Pl. 1, Fig. 4).

2-1. Nymphaeaceae (Pl. 1, Figs. 5-8) Specimens examined: Nuphar advena (Aiton) W.T. Aiton; BGZ 537, BGZ 19963491 The gynoecia of Nymphaeaceae are partially fused to syncarpous ovaries composed of 1-47 carpels arranged in whorls (Endress 2001). The carpels contain three to 400 or more, anatropous (except for Barclaya, which has orthotropous (Schneider 1978)), bitegmic and crassinucellate ovules (Iger- sheim and Endress 1998). Most taxa have an aril except for Nuphar (“aril- loid” in seed) and Barclaya (Hart and Cox 1995, Schneider and Williamson 1993). The micropyle is formed by the annular, unlobed inner integument in Barclaya, Nuphar, Nymphaea or by both integuments in Euryale, Victoria, and Barclaya (Igersheim and Endress 1998, Winter 1993). The outer in- tegument is unlobed, semiannular or sometimes annular (Nuphar, Nym- phaea), or always annular (Barclaya, Euryale, Victoria) (Igersheim and En- dress 1998). The fruits are fleshy berries, 3.5-4.2 cm long and 2.5-2.9 mm broad in examined species (Nuphar advena).

16 The following seed description is based on one species (Nuphar advena) from Nymphaeaceae. Seeds are ovoid in shape, close to yellow brown in colour. Dimensions of seed: length 5.3-7.8 mm; width 3.8-4.7 mm; thickness 3.6-4.2 mm. Exotestal cells in surface view sinous, 52-81 m x 18-31 m (Pl. 1, Fig. 7). Outer sur- face of testa is smooth. Seeds are exotestal with seed coat formed mainly by the testa. Testa is about 200 m thick, composed of one palisade outer layer and an inner sclerotic cell layer. The outer palisade cell layer is 85-90 m high, with simple, minute pores in the anticlinal cell walls. The inner sclerotic cell layer is 105-115 m thick, 3-5 cells deep consisting of more or less trans- versely elongated cells with lumens (Pl. 1, Fig. 8). The inner integument in the ovules consists of two cell layers (Igersheim and Endress 1998). In mature seeds the tegmen is crushed and the number of cell layers cannot be estab- lished. The circular cap, present at the apex, has micropyle in the center. Hi- lum scar extends between micropyle and end of raphe ridge (Pl. 1, Figs. 5-6). Seeds of other Nymphaeaceae have the same principal organization and structure, but there is also a considerable diversity expressed, e.g., in the shape of the exotestal cells and the position of hilum and micropyle in rela- tion to the germination cap (e.g., Collinson 1980). Exotestal cells in surface view of other taxa (e.g., Nymphaea, Ondinea, Victoria) in Nymphaeaceae have strongly undulate cells (Collinson 1980). Caps showing position of micropyle and hilum in different genera have morphological differences in Nymphaeaceae as well as testa structure and seed size (See Figs.1-5 in Col- linson 1980).

17

Plate 1. Figs. 1-4. Fruits and seeds of Amborellaceae. Fig. 1. , at left. Fig. 2. Longitudinal section of fruit with intact seed inside. Fig. 3. Seed in lateral view. Arrow heads indicate micropylar area. Fig. 4. Micropylar area of seed. Figs. 5- 8. Seeds of Nuphar advena (Nymphaeaceae). Fig. 5. Apical view of seed. Arrow heads indicate raphe ridge. Fig. 6. Circular cap showing micropylar area in center and hilum scar between micropylar area and end of raphe ridge. Fig. 7. Outer surface of testa showing sinous cells. Fig. 8. Longitudinal section of seed wall. Scale bars: Figs. 1-3, 5: 1 mm, Figs. 4, 8: 100 μm, Fig. 6: 300 μm, Fig. 7: 50 μm.

18 2-2. Cabombaceae (The family is included in Nymphaeaceae according to APG 2003) The gynoecia of Cabombaceae are composed of (1-) 2-18 (Cabomba (1-) 2-3 (-7) whorled and Brasenia 4-22 probably arranged in whorls) free carpels (Ito 1987, Raciborski 1894, Williamson and Schneider 1993). The carpels contain (1-) 3 (-5) ovules in Cabomba, 1 or 2 in Brasenia. The ovules are anatropous, bitegmic and crassinucellate (Igersheim and Endress 1998). An aril is not present. The micropyle is formed by the annular, unlobed inner integument. The outer integument is semiannular and unlobed (Igersheim and Endress 1998). The fruit is non-fleshy and aggregate; indehiscent or dehiscent with 2-3 seeds per follicle-like fruit in Cabomba and 1-2 seeds per achene-like fruit in Brasenia (Williamson and Schneider 1993). The operculate seeds are endospermic, ovoid in shape, perisperm present with two cotyledons (Watson and Dallwitz 1992, Williamson and Schneider 1993). Collinson (1980) provided a useful survey of the seeds in Nym- phaeaceae and Cabombaceae and the following description is mainly based on this work. Seeds are exotestal. Exotesta cells show distinct undulate anti- clinal walls in surface view. In Brasenia the exotesta is developed as a dis- tinct palisade layer, very similar to that of Illicium and some Schisandraceae. In Cabomba exotestal cells are generally lower and with tubercular protru- sions. Inner layers of testa one (Cabomba) to three (Brasenia) cell layers deep of sclerotic cells. Tegmen is membranous and crushed. Both Cabomba and Brasenia has a distinct germination cap with micropyle and hilum placed more or less in the center.

3. Trimeniaceae (Pl. 2, Figs. 1-4) Specimens examined: Trimenia weinmanniifolia Seem.; A. C. Smith 8355 (S) The gynoecia of Trimeniaceae are composed of one (rarely two) carpel(s) (Endress and Igersheim 1997, Philipson 1993). The carpels contain a single, anatropous, bitegmic and crassinucellate ovule (Endress 2001). The micro- pyle is formed by the lobed outer and inner integuments (Endress and Iger- sheim 1997). The fruit is berry, spherical in shape (Pl. 2, Fig. 1), 4.7-6 mm long and 2.7-3 mm broad. The pedicel is 1-2.2 mm long. The epicarp of the berry has a more or less smooth surface. Seeds one per carpel, more or less ovoid in shape with micropylar area protruding (Pl. 2, Fig. 2). Dimensions of seed: length 3.0-3.5 mm; width 2.1- 2.3 mm; thickness 1.9-2.1 mm. Outer surface of testa is reticulate with finely undulate cell walls (Pl. 2, Figs. 3-4). Testa is about 160-180 m thick. Teg- men consists of three cell layers in the ovule (Endress and Igersheim 1997). In mature seeds the tegmen is crushed and the number of cell layers cannot be established.

19 4. Austrobaileyaceae (Pl. 2, Figs. 5-8) Specimens examined: Austrobaileya scandens C. T. White; B. Gray 1123 (L) The gynoecia of Austrobaileyaceae are apocarpous composed of 10-13 free, spirally arranged carpels (Endress 1993, Igersheim and Endress 1997). Each carpel contains (4-) 8-14 ovules arranged in two longitudinal series along the sutures (Bailey and Swamy 1949, Endress 1980, Takhtajan 1988). The ovules are anatropous, bitegmic, and crassinucellate. The micropyle is formed by the slightly lobed and annular inner integument. The outer in- tegument is unlobed (Igersheim and Endress 1997). The fruits are ellipsoidal or globose fleshy berries, 4.2-7 cm long and 3.1-4 cm broad with a 1.3-2 cm long stalk (pedicel) (Pl. 2, Fig. 5). The epicarp of the berry is orange col- oured and the fleshy endocarp is yellow (Endress 1980, Endress 1993). Seeds three per fruit (Pl. 2, Fig. 6), broadly ovate, lenticular, and chest- nut-like in shape (Pl. 2, Fig. 7), close to orange brown in colour. The seeds are ovate in transverse section. Dimensions of seed: length 2.3-2.5 cm; width 1.6-2.3 cm; thickness 1.3-2.0 cm. The outer surface of seed is ruminated. Seeds are sarcotestal with an outer soft layer and an inner sclerotic layer. Testa is about 850-1150 m thick. The outer layer is composed of a thin cell layer, 20-40 m high. The inner sclerotic cell layer is 810-1110 m thick, 9- 12 cells deep, with simple, minute pores in the anticlinal cell walls. The in- ner integument in the ovules consists of two (or three) cell layers (Igersheim and Endress 1997). In mature seeds the tegmen is crushed and the number of cell layers cannot be established. In lateral view the hilar area is obtuse and the chalazal area is rounded (Pl. 2, Fig. 7). A distinct raphe ridge extends from the hilar area to the chalaza and antiraphe ridge is also present.

20

Plate 2. Figs. 1-4. Fruits and seeds of Trimenia weinmanniifolia (Trimeniaceae). Fig. 1. Fruit, pedicel at right. Fig. 2. Seed in lateral view, micropylar area at right. Fig. 3. Micropylar area of seed in lateral view. Fig. 4. Surface of seed showing reticulation with finely undulate cell walls. Figs. 5-8. Fruits and seeds of Austrobaileyaceae. Fig. 5. Fruit with seeds inside. Fig. 6. Three seeds in apical view. Arrow heads indicate hilar area. Fig. 7. Seed, hilar area at right. Fig. 8. Detail of hilar area. Arrow head indicates hilar area. Scale bars: Figs. 1-3: 1 mm, Fig. 4: 100 μm, Figs. 5-7: 1 cm, Fig. 8: 1.2 mm.

21 Discussion of papers I and II Fossil record There is no reliable fossil record of Amborella. Vesselless fossils de- scribed from India (Jurassic ?), Western North America (Cainozoic), and Greenland (Eocene) (Bailey and Swamy 1948a) are all problematic and have not been convincingly documented. Anatomical features of the secondary xylem (lacking vessels) obviously are indistinguishable from extant Trocho- dendron and Tetracentron (Bailey and Swamy 1948a). There is no fossil record of Austrobaileyaceae and Trimeniaceae. In contrast, Illiciaceae, Schi- sandraceae, and Nymphaeaceae have a fossil record that extends back to the Cretaceous (Jähnichen 1976; , Manchester 1994, Frumin and Friis 1999, Friis et al. 2001; reproductive structures). While seeds from the Clarno Formation of the Eocene of Oregon are very similar to the modern North American Schisandra glabra in size, shape, seed surface, and hilar area, Late Cretaceous seeds from Central Asia with affinities to Illiciaceae differ from modern species by their extremely small size.

Phylogenetic considerations Both seeds of Illiciaceae and Schisandraceae display characteristics that are useful for intrageneric classification (see the hilar rim in Illicium, paper 1, and the seed surface and number of cells forming the mesotesta in Schi- sandraceae, paper 2). The overall appearance of seeds belonging to these two families, however, does not provide any deeper insight into the phylogeny of basal angiosperms in general, i.e., there is no clear phylogenetic signal for a number of key-characters, such as the campylotropous versus anatropous seed organization in Schisandraceae and Illiciaceae, respectively. Another feature, the size of seeds in modern Illiciaceae and Schisandraceae does not correspond to the size classes of seeds reported for the Cretaceous (Tiffney 1984, Frumin and Friis 1999; Eriksson et al. 2000). The predominance of small seeds in the early fossil record of flowering plants also contradicts previous ideas about primitive characters in seeds, e.g. Goldberg (1986), who suggested medium-sized seeds to be primitive within angiosperms. The nature of the ancestral seed type within angiosperms has been a mat- ter of strong debate (Bocquet and Bersier 1960, Boesewinkel and Bouman 1984, Eames 1961). The presence of both anatropous and orthotropous seeds in early fossils of the angiosperms (Friis et al. 1997) and in extant ANITA members (see above), together with the lack of a solid phylogenetic frame- work for seed plants, make it impossible for the present to say whether one or the other type should be ancestral within angiosperms. More interesting is the fact that basal angiosperm seeds are diverse in structure and organization. This is not surprising given the long history of

22 the ANITA grade and the enormous diversity in habit and reproductive biol- ogy exhibited by its various members.

Biogeography Many extant basal angiosperms show disjunct distributions in both Asia and eastern North America, e.g., Illiciaceae and Schisandraceae. The fossil re- cord supports the antiquity of ANITA lineages with unambiguous Early and mid-Cretaceous finds of Nymphaeales (Friis et al. 2001) and Illiciaceae (Frumin and Friis 1999). The fossil record of Illiciaceae and Schisandraceae further demonstrates that this ANITA family also had a more extensive geo- graphical distribution in earlier geographical periods. Fruit, leaf and seed fossils of different periods during the Late Cretaceous and Early Cainozoic have been described from Europe, northwestern Asia, and North America (Frumin and Friis 1999, Mai 1970, Tiffney and Barghoorn 1979).

23 Paper III

In this paper we investigated seed morphology in the genus Lysimachia and in six additional genera (, , Asterolinon, Glaux, Pel- letiera, Trientalis), which are considered to be closely related to, or are placed within Lysimachia in previous molecular studies (Hao et al. 2004, Manns and Anderberg 2005, Anderberg et al. 2007). The results in paper III show that three major types of seed shape can be recognized: 1) sectoroid, 2) polyhedral, and 3) coarsely rugose with concave hilar area. Most species correspond to the first type and have an obliquely sectoroid seed shape. The seeds are dorsiventrally and/or laterally flattened. Taxa with this type of seeds include the genus Anagallis, Glaux maritima, Asterolinon adoënse as well as most Lysimachia species. In addition, in sec- toroid and polyhedral seeds the edges may be keeled or winged. The third seed shape type (coarsely rugose with concave hilar area) is only present in Asterolinon linum-stellatum and in the two species of genus Pelletiera. In their seed coat structure, most species are characterized by a thin and continuous outer seed coat layer that is closely adhering to the inner seed coat. In a few species (Trientalis europaea, , L. thyrsi- flora, and L. terrestris), however, the outer layer of the seed coat is sponge- like. Seed surface patterns can be divided into six main types: 1) reticulate, 2) tuberculate, 3) vesiculose, 4) colliculate, 5) undulate, and 6) poroid- alveolate. Seed surface patterns are mostly congruent with molecular phy- logenetic relationships. A reticulate surface pattern is diagnostic of, e.g., Lysimachia subgenera Palladia and Hawaiian Lysimachiopsis. Mapping seed characters onto a recent phylogenetic tree from Anderberg et al. (2007) (Fig. 5), reveals that they provide potentially synapomorphic character states for various subclades of Lysimachia (Fig. 6). Rugose seed shape and vesiculose seed coat surface are synapomorphic for the clade comprising the genus Pelletiera and Asterolinon linum-stellatum (clade III). A clade with Lysimachia vulgaris, L. thyrsiflora, and L. terrestris is charac- terized by a poroid-alveolate seed coat surface and a sponge-like outer seed coat layer (clade IV). Both of these character states have apparently evolved in parallel in Trientalis europaea (clade I) and along the branch leading to the clade with the three Lysimachia species. We also discuss possible habitat factors that may have favoured the independent evolution of particular seed types, such as winged seeds in various lineages.

24 Our conclusion is that seed morphology in Lysimachia and related taxa clearly is of systematic interest. In particular, seed shape and the structure of the seed coat provide potentially synapomorphic character states for various subclades of Lysimachia.

Ardisiandra wettsteinii Outgroup Trientalis europaea Trientalis - L. andina - L. ciliata 100 Subg. Lysimachia A 62 L. quadriflora L. hybrida L. punctata Subg. Lysimachia B

Anagallis minima 100 55 Anagallis serpens Anagallis p.p. 100 Anagallis pumila 93 - Anagallis tenuicaulis Pelletiera verna 100 85 Asterolinon linum-stellatum Pelletiera + Asterolinon p.p. Pelletiera wildpretii 100 100 Anagallis p.p. 94 Anagallis monellii 99 Asterolinon adoense Asterolinon p.p. L. nemorum Subg. Lysimachia C 52

92 L. laxa Subg. Idiophyton L. insignis

L. vulgaris Subg. Lysimachia E 99 100 L. thyrsiflora Subg. Naumburgia L. terrestris Subg. Lysimachia E

L. atropurpurea

91 L. mauritiana

L. ephemerum 63 L. ruhmeriana Subg. Palladia

L. lobelioides

80 - L. clethroides L. candida

Glaux maritima Glaux

- L. alfredii

L. paridiformis 91 L. christinae

L. fordiana Subg. Lysimachia F 88 L. phyllocephala - - L. remota L. japonica Figure 5. Simplified phylogenetic tree redrawn from the phylogenetic analysis of ndhF sequences by Anderberg et al. (2007). Note that this tree only includes species studied in the present paper, all others were pruned from the original tree. A clade (subg. Lysimachia A) including T. europaea is sister to all remaining Lysimachia species. Jackknife support values are given above branches.

25

Figure 6. Seed morphological characters mapped onto the simplified (only species included in the present study) phylogenetic tree from Anderberg et al.’s ndhF se- quences (2007). Dimorphic character states are indicated with a shaded square ( ). Major clades referred to in the text are numbered I–VI. A Seed shape. B Seed coat surface. C Sponge-like outer layer. D Wing

26 Paper IV

Paper IV focuses on the relationships of the endemic Hawaiian Lysimachia species constituting L. subgenus Lysimachiopsis. There are 16 endemic spe- cies of Lysimachia in Hawaii, which show considerable morphological variation, especially in leaf and structure (Marr and Bohm 1997). However, previous taxonomic treatments indicate that the delimitation of taxa is problematic, especially for the four of L. remyi (ssp. re- myi, caliginis, kipahuluensis, subherbacea) and the species of the L. hille- brandii complex (i.e. L. hillebrandii, L. waianaeensis, L. ovoidea), which were recognized as distinct by Marr and Bohm (1997). We used nrDNA (ETS and ITS) and cpDNA (rpl16, rpl20-rps12, rps16, trnH-psbA, trnS-G) sequence data in order to reconstruct phylogenetic rela- tionships. In general, nrDNA data sets contain more informative data and provide better resolution in phylogenetic trees than data sets from chloro- plast markers. A phylogenetic analysis based on the combined data set using all nrDNA and cpDNA sequences provides new insights into the relation- ships within subgenus Lysimachiopsis, especially within the problematic Lysimachia hillebrandii/L. remyi complex (Fig. 7). The result from the mo- lecular analysis provides support for the following clades: (1) L. hillebrandii and L. waianaeensis (clade D, 55% jackknife support: JS), (2) L. remyi ssp. remyi, L. maxima, and L. remyi ssp. subherbacea (clade E, 86% JS), (3) L. remyi ssp. caliginis and L. remyi ssp. kipahuluensis (clade G, 93% JS), and (4) L. glutinosa, L. scopulensis, and L. kalalauensis (clade H, 81% JS). The phylogenetic relationships among these taxa are largely congruent with the biogeographical distribution of the species in the Hawaiian Islands. Our re- sults also indicate that earlier taxonomic treatments of the group need to be partially revised in order to reflect phylogenetic relationships.

27

L. mauritiana Outgroup (Coastal E. Asia, Hawaiian Is., Oceania) 90 L. iniki (1) (Kaua'i) A L. iniki (2) (Kaua'i)

L. filifolia (Kaua'i, O'ahu) L. filifolia x glutinosa (Kaua'i ?)

82 55 L. hillebrandii (O'ahu, Moloka'i) B D L. waianaeensis (O'ahu)

L. remyi ssp. remyi (Maui, Lāna'i) 86 E L. maxima (Moloka'i) Subg. 86 F L. remyi ssp. subherbacea Lysimachiopsis (Moloka'i, O'ahu) L. daphnoides (Kaua'i) L. ovoidea (Kaua'i)

L. remyi ssp. caliginis (Maui) 93 66 G 87 L. remyi ssp. kipahuluensis (1) (Maui) C L. remyi ssp. kipahuluensis (2) (Maui) 73 L. glutinosa (1) (Kaua'i) L. glutinosa (2) (Kaua'i) 81 H L. scopulensis (1) (Kaua'i) 54 L. scopulensis (2) (Kaua'i) L. glutinosa x kalalauensis (Kaua'i) 92 59 L. kalalauensis (1) (Kaua'i) L. kalalauensis (2) (Kaua'i) Figure 7. Jackknife tree obtained from Xac analysis of combined nrDNA (ETS, ITS) and cpDNA (rpl16, rpl20-rps12, rps16, trnH-psbA, trnS-G) data sets. Jackknife support values ( 50%) are indicated above the nodes. Clades referred to in the text are designated by the letters A-H.

28 Sammanfattning (Swedish summary)

Fröväxter och frön Fröväxterna är den största och mest differentierade gruppen bland växterna. De kännetecknas av att ha just frön, och uppkomsten av fröet innebar uppen- barligen en revolution i växternas utvecklingshistoria. Frön är reproduktions- och spridningsorgan, som också ger skydd åt och förser det växande embryot med näring. Frön kan tåla extrema förhållanden, både värme och kyla, och kan ligga många år i vila i jorden innan de gror. Den största gruppen inom fröväxterna är blomväxterna, som funnits på jorden åtminstone sen tidig krita. De kännetecknas bland annat av att fröäm- nena, de unga fröna, är inneslutna i ett fruktämne som utvecklas till en frukt. En ytterligare specialisering hos blomväxterna, jämfört med övriga fröväxter, är att frönas näringsvävnad, endospermet, enbart bildas i fröämnen som bli- vit befruktade. Detta är en energibesparande mekanism och kanske en av faktorerna bakom blomväxternas framgång. Frön är ett genomgående tema i den här avhandlingen, där frömorfologi studeras, i ett fylogenetiskt perspektiv, i två olika delar av blomväxtsystemet. Artikel I och II fokuserar på de basala grupper av angiospermer som brukar sammanfattas under beteckningen ANITA, medan artiklarna III och IV be- handlar frömorfologi och fylogeni hos en grupp asterider, nämligen Lysima- chia och närstående taxa i familjen Myrsinaceae.

ANITA Det har länge spekulerats om vilken av de nulevande grupperna av blomväx- ter som sitter vid eller nära basen av blomväxternas släktträd. Genom fylo- genetiska analyser baserade på molekylära data börjar nu bilden klarna och några återkommande basala grupper i dessa analyser är Amborella, ett släkte med en enda art som utgör familjen Amborellaceae, Nymphaeaceae, näck- rosväxterna, som bildar en grupp tillsammans med Cabombaceae, Illiciaceae, stjärnanisväxterna, som bildar en grupp tillsammans med Schisandraceae, Trimeniaceae och Austrobaileyaceae. De här grupperna av basala blomväx- ter brukar ofta betecknas ANITA efter de initialer som markerats ovan. ANITA är ingen naturlig (monofyletisk) grupp, utan består av en samling separata utvecklingslinjer nära blomväxternas rot.

29 Nyligen har ytterligare en grupp visat sig höra till ANITA, nämligen den lilla familjen Hydatellaceae, som tidigare trotts vara monokotyledoner. Hy- datellaceae består av små akvatiska gräslika örter som, mycket oväntat, tycks ha en fylogenetisk position nära näckrosväxterna.

Frömorfologi hos ANITA Bland de fossil av växter som upptäckts och upptäcks finns många välbeva- rade frön, som kan ha mer eller mindre stora likheter med frön från nulevan- de grupper. Artiklarna I och II, liksom avsnitten på sid. 17–22 i avhandling- en, behandlar frömorfologi hos olika grupper inom ANITA. Även blad är ofta bevarade som fossil och i artikel I och II studeras, förutom frön, även epidermis från blad hos nulevande grupper inom ANITA. Artikel I fokuserar på släktet Illicium, en grupp med omkring 35 arter i Asien och Amerika, den mest kända I. anisatum, stjärnanis. En tidigare mo- lekylär studie indikerade en uppdelning i en asiatisk och en amerikansk gren inom Illicium. Vi kunde nu visa att den uppdelningen också stöds av struktu- rer i fröna hos dessa växter. Fossila frön (och blad) som associerats med Illicium utvärderas också med hjälp av dessa nya data. I artikel II studeras frömorfologi och bladepidermis hos de två närstående släktena Kadsura (i Asien) och Schisandra (i Asien och Amerika) i familjen Schisandraceae. Här adderades våra data till en redan existerande datamatris med morfologiska data. En tidigare fylogenetisk analys baserad på dessa data stödde den rådande uppfattningen att Kadsura och Schisandra är mono- fyletiska grupper, men våra nya data gav inget ökat stöd för detta. I stället kunde vi påvisa ett parvis släktskap mellan vissa arter inom dessa grupper. Fossila frön av Schisandra från Nordamerika visar likheter med en nulevan- de nordamerikansk art, medan fossila frön från Europa överensstämmer bäst med nulevande arter i Asien.

Frömorfologi hos Lysimachia Lysimachia hör till asteriderna, en mycket stor grupp som sitter i toppen av blomväxternas släktträd och omfattar omkring 1/3 av alla blomväxter. Hit hör nästan alla grupper med sammanväxta kronblad. Lysimachia har tradi- tionellt förts till familjen Primulaceae, men med en snävare familjeavgräns- ning placeras den numera ofta i Myrsinaceae i ordningen Ericales. Till Lysi- machia hör välkända svenska växter som videört och topplösa. Artikel III fokuserar på frömorfologi hos Lysimachia och de närstående släktena Anagallis (med bl. a. rödmire), Ardisiandra, Asterolinon, Glaux (strandkrypa), Pelletiera och Trientalis (skogsstjärna). Molekylära studier indikerar att Lysimachia med sin nuvarande avgränsning inte är en monofy- letisk grupp. För att den ska bli det måste dessa närstående släkten, helt eller

30 delvis, också inkluderas. Det är därför viktigt att samtliga dessa släkten är med i den frömorfologiska studien. Våra resultat visar att frömorfologiska karaktärer i flera fall stöder olika undergrupper inom Lysimachia i vid mening. Fröskal med retikulat ytmöns- ter, till exempel, karaktäriserar både undersläktet Palladia och dess förmo- dade systergrupp, undersläktet Lysimachiopsis, en grupp arter som endast växer på Hawaii.

Fylogeni inom Lysimachia på Hawaii I artikel IV inkluderas de flesta kända arterna inom undersläktet Lysimachi- opsis, inklusive några underarter, i en fylogenetisk analys baserad på DNA sekvenser både från kärngenomet och kloroplastgenomet. Kloroplastsekven- serna innehålller genomgående mindre fylogenetiskt användbar information än kärngenomet, men analyser baserad på ett kombinerat dataset med samt- liga sekvenser ger ett ganska väl upplöst släktträd. De olika grenarna omfat- tar ofta arter och underarter med utbredningen begränsad till vissa öar inom Hawaii-arkipelagen. Resultaten indikerar också att systematiken inom vissa taxonomiskt problematiska grupper behöver revideras.

31 Acknowledgements

The research work presented in the thesis has been carried out at the De- partments of Palaeobotany and Phanerogamic Botany, Swedish Museum of Natural History and Department of Systematic Biology (previously System- atic Botany), Evolutionary Biology Centre (EBC), Uppsala University. First of all, I would like to thank to my supervisor Professor Arne A. An- derberg, for his excellent academic guidance, endless encouragement, and trust. I have enjoyed very much working on the Lysimachia project! Special thanks are also given to my co-supervisor Dr. Jürg Schönenberger for his kind help with manuscripts and many suggestions. I am very grateful to my official supervisor Professor Mats Thulin in Uppsala University for reading manuscripts and help with so many things in the department. I would like to thank Professor Sandra Baldauf, Professor Leif Tibell and Dr. Petra Korall for constructive comments on the manuscript of paper IV. I am grateful to all members in Phanerogamic and Cryptogamic Botany Departments, Swedish Museum of Natural History. Special thanks to Ulf Swenson, Anna-Lena Anderberg, Johannes Lundberg, Jens Klackenberg, Mia Ehn, Lennart Stenberg, Thomas Karlsson, Dick Andersson, Maria Back- lund, Anvar Sheikhahmadi, Gunilla Dahlerus Lehman, Ulla-Karin Skärlund, and Erik Emanuelsson for their kind help and many suggestions. Many thanks go to all former and present members in Molecular Sys- tematics Laboratory, especially Mattias Myrenås, Pia Eldenäs, Ida Trift, Martin Irestedt, James S. Farris, and Mari Källersjö for their kind help and suggestions. I would like to thank people in Botany Department, Stockholm Univer- sity, special thanks to Ulrika Manns, Markus Englund for their kind help and suggestions. I would like to express my sincere gratitude to my previous supervisor Professor Else Marie Friis for giving me an opportunity to do my PhD study (basal angiosperms project), initiating the idea of my PhD project, academic guidance, encouragement, patience, and positive and constructive discus- sions. I would like to thank my previous co-supervisor Professor Birgitta Bremer for accepting me as a PhD student at Uppsala University, very fruitful dis- cussion, and helpful comments on my manuscripts.

32 Thanks go to Dr. Thomas Denk as both my dear colleague and unofficial supervisor. He is a ‘super hero’ in the palaeobotany department and a lot of help to many people including me. I would like to express my deepest thank to Thomas for his encouragement, friendship, constructive criticism, unlim- ited ideas, and help with various other things. I am very much grateful to Drs Jürg Schönenberger and Maria von Bal- thazar, for their kindness, supplying material for my study, and many valu- able suggestions. Previous and present members of the Palaeobotany Department at the Swedish Museum of Natural History are thanked for their kindness and help. The more than warm atmosphere at the department has been a firm founda- tion for my life in Sweden. I would like to thank Kamlesh Khullar for help with many practical things. Special thanks to Dr. David Cantrill for English correction in my manuscripts; Catarina Rydin for helpful discussions and suggestions. Yvonne Arremo for help with SEM and photography, for deli- cious bakery for afternoon coffee breaks; Ove Johansson, Per-Olof Haglund, Lars Imby, Sandra May, Hervé Sauquet, and Caroline Strömberg for help with various things. Thanks go to Dr. Ki Andersson and other members at the Department of Palaeozoology, Swedish Museum of Natural History for their kind help. I am grateful to all people in Systematic Botany department, EBC, Uppsa- la University for their kind help and encouragement. Special thanks to Björn-Axel Beier, Per Kornhall, Magnus Popp, Jesper Kårehed, Annika Vinnersten, Per Erixon, Dick Andersson, Maria Backlund, Niklas Wikström, Elisabeth Långström, Bengt Oxelman, Sylvain Razafimandimbison, Katarina Andreasen, Hugo De Boer, Anneleen Kool, Magnus Lidén, Kristina Articus, Henrik Lantz, Inga Hedberg, Frida Eggens, Cajsa Anderson, Ulla Heden- quist, Catarina Ekenäs, Anja Rautenberg, Sunniva Aagaard, and Johannes Lundberg for their kind help and many suggestions. I would like to thank Professor Suk-Pyo Hong, who introduced me to bot- any and who has always encouraged me in various ways. Thanks also to many Korean people and friends in Sweden for their con- cerns and support, especially Hyun-Duk Kim and Kyung-Sil Choo for their encouragement. Last, but not least, I would like to thank my family for their endless love and support across the Eurasian continent. I hope this thesis would be a small gift to my mother in Korea and father in heaven. I always think of you and bear in mind your favorite motto, “Always be sincere and diligent”.

33 References

Anderberg A. A., Ståhl B. 1995. Phylogenetic interrelationships in the order Primu- lales, with special emphasis on the family circumscriptions. Canadian Journal of Botany 73: 1699-1730. Anderberg A. A., Ståhl B., Källersjö M. 1998. Phylogenetic relationships in the Primulales inferred from rbcL sequence data. Plant Systematics and Evolution 211: 93-102. Anderberg A. A., Rydin C., Källersjö M. 2002. Phylogenetic relationships in the order Ericales s.l.: analyses of molecular data from five genes from the plastid and mitochondrial genomes. Am. J. Bot. 89: 677-687. Anderberg A. A., Manns U., Källersjö M. 2007. Phylogeny and floral evolution of the Lysimachieae (Ericales, Myrsinaceae): evidence from ndhF sequence data. Willdenowia 37: 407-421. APG 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Botanical Journal of the Lin- nean Society 141: 399-436. Arber E. A. N., Parkin J. 1907. On the origin of angiosperms. Journal of the Linnean Society Botany 38: 29-80. Bailey I. W., Swamy B. G. L. 1948a. Amborella trichopoda Baill., a new morpho- logical type of vesselless . Journal of the Arnold Arboretum 29: 215. Bailey I. W., Swamy B. G. L. 1949. The morphology and relationship of Austrobai- leya. Journal of the Arnold Arboretum 30: 211-226. Bocquet G., Bersier J. D. 1960. La valeur systématique de l'ovule: Développments tératologiques. Arch. Sc. (Genève) - Archives Des Sciences 13: 475-496. Boesewinkel F. D., Bouman F. 1984. The seed: structure. In: Johri B. M., Ed. Em- bryology of angiosperms. Springer, Berlin. Bouman F. 1974. Developmental studies of the ovule, integuments and seed in some angiosperms. Ph. D. thesis, Univ. of Amsterdam, Los, Naarden Buxbaum F. 1961. Vorläufige Untersuchungen über Umfang, systematische Stel- lung und Gliederung der (Centrospermae). Beiträge zur Biolo- gie der Pflanzen 36: 1-56. Chase M. W., Soltis D. E., Olmstead R. G., Morgan D., Les D. H., Mishler B. D., Duvall M. R., A. P. R., Hills H. G., Qiu Y.-L., Kron K. A., Rettig J. H., Conti E., Palmer J. D., Manhart J. R., Sytsma K. J., Michaels H. J., Kress W. J., Karol K. G., Clark W. D., Hedrén M., Gaut B. S., Jansen R. K., Kim K.-J., Wimpee C. F., Smith J. F., Furnier G. R., Strauss S. H., Xiang Q.-Y., Plunkett G. M., Soltis P. S., Swensen S. M., Williams S. E., Gadek P. A., Quinn C. J., Eguiarte L. E., Golenberg G., Learn G. H. J., Graham S. W., Barrett S. C. H., Dayanandan S., Albert V. A. 1993. Phylogenetics of seed plants: an analysis of nucleotide se- quences from the plastid gene rbcL. Annals of the Missouri Botanical Garden 80: 528-580. Chen F.-H., Hu C.-M. 1979. Taxonomic and phytogeographic studies on Chinese species of Lysimachia. Acta Phytotaxonomica Sinica 17: 21-53.

34 Collinson M. E. 1980. Recent and Tertiary seeds of the Nymphaeaceae sensu lato with a revision of Brasenia ovula (Brong.) Reid and Chandler. Annals of Bot- any 46: 603-632. Corner E. J. H. 1976. The seeds of . Cambridge University Press, Cam- bridge. Cronquist A. 1981. An integrated system of classification of flowering plants. Co- lumbia University Press, New York. Donoghue M. J., Doyle J. A. 1989. Phylogenetic analysis of angiosperms and the relationships of Hamamelidae. In: Crane P. R., Blackmore S., (eds.) Evolution, systematics , and fossil history of the Hamamelidae, Vol. 1. Introduction and 'lower' Hamamelidae. Clarendon Press, Oxford, pp. 17-45. Doyle J. A. 1994. The origin of the angiosperm : A phylogenetic perspective. Plant Systematics and Evolution (Suppl.) 8: 7-29. Eames A. J. 1961. Morphology of the angiosperms. McGraw-Hill, New York. Endress P. K. 1980. The reproductive structure and systematic position of the Aus- trobaileyaceae. Botanische Jahrbücher für Systematik 101: 393-433. Endress P. K. 1987. The Chloranthaceae: reproductive structures and phylogenetic position. Botanische Jahrbücher für Systematik, Pflanzengeschichte und Pflan- zengeographie 109: 153-226. Endress P. K. 1993. Austrobaileyaceae. In: Kubitzki K., Rohwer J. G., Bittrich V., (eds.) The families and genera of vascular plants. II Flowering plants - dicotyle- dons. Magnoliid, hamamelid and caryophyllid families. Springer, Berlin, pp. 138-140. Endress P. K. 1994. Floral structure and evolution of primitive angiosperms: recent advances. Plant Systematics and Evolution 192: 79-97. Endress P. K. 2001. The flowers in extant basal angiosperms and inferences on an- cestral flowers. International Journal of Plant Science 162: 1111-1140. Endress P. K., Igersheim A. 1997. diversity and systematics of the Lau- rales. Botanical Journal of the Linnean Society 125: 93-168. Endress P. K., Igersheim A. 2000. The reproductive structures of the basal angio- sperm Amborella trichopoda (Amborellaceae). International Journal of Plant Science 161 (6 suppl.): S237-S248. Eriksson O., Friis, E. M., Löfgren, P. 2000. Seed size, fruit size and dispersal sys- tems in angiosperms from the Early Cretaceous to the Late Tertiary. The Ameri- can Naturalist 156: 47-58. Friis E. M., Crane P. R., Pedersen K. R. 1997. Fossil history of magnoliid angio- sperms. In: Iwatsuki K., Raven P. H., (eds.) Evolution and diversification of land plants. Springer Verlag, Tokyo, Berlin, Heidelberg, New York, pp. 121- 156. Friis E. M., Pedersen K. R., Crane P. R. 2001. Fosssil evidence of water lilies (Nymphaeales) in the Early Cretaceous. Nature 410: 357-360. Friis E. M., Crane P. 2007. Botany: New home for tiny aquatics. Nature 446: 269- 270. Frumin S., Friis E. M. 1999. Magnoliid reproductive organs from the Cenomanian- Turonian of north-western Kazakhstan: Magnoliaceae and Illiciaceae. Plant Sys- tematics and Evolution 216: 265-288. Fukuhara T. 1999. Seed and funicle morphology of Fumariaceae-Fumarioideae: systematic implication and evolutionary patterns. International Journal of Plant Science 160: 151-180. Goldberg A. 1986. Classification, evolution, and phylogeny of the families of di- cotyledons. Smithsonian Contributions to Botany no. 58: 1-314.

35 Hao G., Chye M.-L., Saunders R. M. K. 2001. A phylogenetic analysis of the Schi- sandraceae based on morphology and nuclear ribosomal ITS sequences. Botani- cal Journal of the Linnean Society 135: 401-411. Hao G., Saunders R. M. K., Chye M.-L. 2000. A phylogenetic analysis of the Il- liciaceae based on sequences of internal transcribed spacers (ITS) of nuclear ri- bosomal DNA. Plant Systematics and Evolution 223: 81-90. Hao G., Yuan Y.-M., Hu C.-M., Ge X.-J., Zhao N.-X. 2004. Molecular phylogeny of Lysimachia (Myrsinaceae) based on chloroplast trnL-F and nuclear ribosomal ITS sequences. Molecular Phylogenetics and Evolution 31: 323-339. Hart K. H., Cox P. A. 1995. Dispersal ecology of Nuphar luteum (L.) Sibthorp & Smith: abiotic seed dispersal mechanisms. Botanical Journal of the Linnean So- ciety 119: 87-100. Hu C.-M., Kelso S. 1996. Primulaceae. In: Wu Z-Y, Raven PH (eds) . vol 15 Science Press, Beijing, China and Missouri Botanical Garden, St. Louis pp. 39-189. Igersheim A., Endress P. K. 1997. Gynoecium diversity and systematics of the Magnoliales and winteroids. Botanical Journal of the Linnean Society 124: 213- 271. Igersheim A., Endress P. K. 1998. Gynoecium diversity and systematics of the pa- leoherbs. Botanical Journal of the Linnean Society 127: 289-370. Ingrouille M. 1992. Diversity and evolution of land plants. Chapman & Hall, Lon- don. Ito M. 1987. Phylogenetic systematics of the Nymphaeales. Botanical Magazine, Tokyo 100: 17-35. Johri B. M., Ambegaokar K. B., Srivastava P. S. 1992. Comparative embryology of angiosperms. Berlin, New York, Springer-Verlag. Jähnichen H. 1976. Schisandraceae und Illiciaceae als holarktische und subtropische Florenelemente im mitteleuropäischen Alttertiär. Abhandlungen des Zentralen Geologischen Institus Paläontologische Abhandlungen 26: 151-197. Keng H. 1993. Illiciaceae. In: Kubitzki K., Rohwer J. G., Bittrich V., (eds.) The families and genera of vascular plants. II Flowering plants - Dicotyledons. Mag- noliid, hamamelid and caryophyllid families. Springer, Berlin, pp. 344-347. Kubitzki K. 1993. Introduction. In: Kubitzki K., Rohwer J. G., Bittrich V., (eds.) The families and genera of vascular plants. II Flowering plants - dicotyledons. Magnoliid, hamamelid and caryophyllid families. Springer, Berlin, pp. 1-12. Källersjö M., Bergqvist G., Anderberg A. A. 2000. Generic realignment in primu- loid families of the Ericales s.l.: a phylogenetic analysis based on DNA se- quences from three chloroplast genes and morphology. American Journal of Botany 87: 1325-1341. Lin Q. 1997. Systematics and evolution of the family Illiciaceae. Ph.D. thesis, For- estry College, South China Agric. Univ., Guangzhou, China Mai D. H. 1970. Subtropische Elemente im europäischen Tertiär I. Die Gattungen Gironniera, Sarcococca, Illicium, Evodia, Ilex, Mastixia, Alangium, und . Paläontologische Abhandlungen, Abteilung B.: Paläobo- tanik 3: 441-503. Manchester S. R. 1994. Fruits and seeds of the Middle Eocene Nut Beds Flora, Clarno Formation, Oregon. Palaeontographica Americana 58: 1-205. Manns U., Anderberg A. A. 2005. Molecular phylogeny of Anagallis (Myrsinaceae) based on ITS, trnL-F, and ndhF sequence data. International Journal of Plant Science 166: 1019-1028 Marr K. L., Bohm B. A. 1997. A taxonomic revision of the endemic Hawaiian Lysi- machia (Primulaceae) including three new species. Pacific Science 51: 254-287.

36 Netolitzky F. 1926. Anatomie der Angiospermen-Samen. Borntraeger, Berlin. Oh I.-C., Denk T., Friis E. M. 2003. Evolution of Illicium (Illiciaceae): Mapping morphological characters on the molecular tree. Plant Systematics and Evolu- tion 240: 175-209. Philipson W. R. 1993. Trimeniaceae. In: Kubitzki K., Rohwer J. G., Bittrich V., (eds.) The families and genera of vascular plants. II Flowering plants - dicotyle- dons. Magnoliid, hamamelid and caryophyllid families. Springer, Berlin, pp. 596-599. Qiu Y. L., Chase M. W., H. L. D., Parks C. R. 1993. Molecular phylogenetics of the Magnoliidae: Cladistic analyses of nucleotide sequences of the plastid gene rbcL. Annals of the Missouri Botanical Garden 80: 587-606. Qiu Y.-L., Lee J., Bernasconi-Quadroni F., Soltis D. E., Soltis P. S., Zanis M., Zimmer E. A., Chen Z., Savolainen V., Chase M. W. 1999. The earliest angio- sperms: evidence from mitochondrial, plastid and nuclear genomes. Nature 402: 404-407. Qiu Y.-L., Lee J., Bernasconi-Quadroni F., Soltis D. E., Soltis P. S., Zanis M., Zimmer E. A., Chen Z., Savolainen V., Chase M. W. 2000. Phylogeny of basal angiosperms: analyses of five genes from three genomes. International Journal of Plant Science 161(6 suppl.): S3-S27. Raciborski M. 1894. Die Morphologie der Cabombeen und Nymphaeaceen. Flora 78: 244-279. Rehder A. 1940. Manual of cultivated trees and shrubs hardy in North America : exclusive of the subtropical and warmer temperate regions. Macmillan, New York. Saarela J. M., Rai H. S., Doyle J. A., Endress, P. K., Mathews S., Marchant A. D., Briggs B. G., Graham S. W. 2007. Hydatellaceae identified as a new branch near the base of the angiosperm phylogenetic tree. Nature 446: 312-315. Saunders R. M. K. 1998. Monograph of Kadsura (Schisandraceae). Systematic Bot- any Monographs 54: 1-106. Saunders R. M. K. 2000. Monograph of Schisandra (Schisandraceae). Systematic Botany Monographs 58: 1-146. Schneider E. L. 1978. Morphological studies of the Nymphaeaceae. IX. The seeds of Barclaya longifolia Wall. Botanical Gazette 139: 223-230. Schneider E. l., Williamson P. S. 1993. Nymphaeaceae. In: Kubitzki K., Rohwer J. G., Bittrich V., (eds.) The families and genera of vascular plants. II Flowering plants - dicotyledons. Magnoliid, hamamelid and caryophyllid families. Sprin- ger, Berlin, pp. 486-493. Smith A. C. 1947. The families Illiciaceae and Schisandraceae. Sargentia 7: 1-224. Sokoloff D. D., Remizowa M. V., Macfarlane T. D., Rudall P. J. 2008. Classifica- tion of the early-divergent angiosperm family Hydatellaceae: one genus instead of two, four new species and sexual dimorphism in dioecious taxa. Taxon 57: 179-200. Soltis D. E., Soltis P. S., Chase M. W., Mort M. E., Albach D. C., Zanis M., Savo- lainen V., Hahn W. H., Hoot S. B., Fay M. F., Axtell M., Swensen S. M., Prince L. M., Kress W. J., Nixon K. C., Farris J. S. 2000. Angiosperm phylogeny in- ferred from 18S rDNA, rbcL, and atpB sequences. Botanical Journal of the Lin- nean Society 133: 381-461. Soltis P. S., Soltis D. E., Chase M. W. 1999. Angiosperm phylogeny inferred from multiple genes as a tool for comparative biology. Nature 402: 402-404. Ståhl B., Anderberg A. A. 2004. Mysinaceae. In: Kubitzki K. (ed.) The families and genera of vascular plants. VI Flowering plants - Dicotyledons. , Ox- alidales, , , Ericales. Springer, Berlin, pp. 266-281.

37 Takhtajan A. 1988. Anatomia seminum comparativa. Vol. 2. Dicotyledons: Magno- liidae, Ranunculidae, pp. 1-256. Izd. Nauka, Leningrad. Takhtajan A. 1997. Diversity and classification of flowering plants. Columbia Uni- versity Press, New York. Tiffney B. H., Barghoorn E. S. 1979. Flora of the Brandon Lignite IV. Illiciaceae. American Journal of Botany 66: 321-329. Tiffney B. H. 1984. Seed size, dispersal syndromes, and the rise of the angiosperms: Evidence and hypothesis. Annals of the Missouri Botanical Garden 71: 551-576. Watson L., Dallwitz M. J. (1992 onwards) The Families of Flowering Plants: De- scriptions, Illustrations, Identification, and Information Retrieval. Version: 25th November 2008. http://delta-intkey.com (last accessed 18th March 2009). Werker E. 1997. Seed anatomy. Gebrüder Borntraeger, Berlin, Stuttgart. Williams J. H., Friedman W. E. 2002. Identification of diploid endosperm in an early angiosperm lineage. Nature 415: 522-526. Williamson P. S., Schneider E. I. 1993. Cabombaceae. In: Kubitzki K., Rohwer J. G., Bittrich V., (eds.) The families and genera of vascular plants. II Flowering plants - dicotyledons. Magnoliid, hamamelid and caryophyllid families. Sprin- ger, Berlin, pp. 157-161. Winter A. N. 1993. Some aspects of the reproductive biology of Hydrostemma lon- gifolium (Barclaya longifolia) (Barclayaceae). Botaniceskij Zhurnal, St. Peters- burg 78: 69-83. Zanis M. J., Soltis D. E., Soltis P. S., Mathews S., Donoghue M. J. 2002. The root of the angiosperms revisited. Proceedings of the National Academy of Science 99: 6848-6853. Zanis M. J., Soltis P. S., Qiu Y. L., Zimmer E., Soltis D. E. 2003. Phylogenetic analysis and perianth evolution in basal angiosperms. Annals of the Missouri Botanical Garden 90: 129-150.

38

 -   -                 .  */0 1 2 ) 1  /

    1  2 ) 1   /3-  -   3 ))  ) 1 )+ 1  4 15 1     6   7 5    + 35   )     +)        )  0    ) 1-  0  1  2 ) 1  /4 8   9 ) 3$,,#3  5 )+ )   :  ) 1-  0   1  2 ) 1  /;4<

        0  +) *)+  4))4    ) *+**))* ",,'$