Evolutionary biology is dependent on accurate phylogenies. In this thesis two branch support methods, Bayesian posterior probablities and bootstrap frequencies, were evaluated with simulated data and empirical data from the chloroplast genome. Bayesian inference was found to be more powerful and less conservative than maximum likelihood bootstrapping, but considerably more sensitive to choice of parameters. Bayesian inference increased in power when data were underparameterized, but the associated increase in type I error was comparatively larger. The chloroplast DNA phylogeny of the tribe Sileneae (Caryophyllaceae) was inferred by analysis of 33,149 aligned nucleotide bases representing 24 taxa. The position of the SW Anatolian taxa Silene cryptoneura and S. sordida strongly disagreed with previous studies on nuclear DNA sequence data, and indicate a possible case of homoploid hybrid origin. Silene atocioides and S. aegyptiaca formed a sister group to Lychnis and remaining Silene, thus suggesting that Silene may be paraphyletic, despite recent revisions based on molecular data. Several nodes in the phylogeny remained poorly supported, despite large amounts of data. Additional sequence sampling is not expected to solve this problem. The main reason for poor resolution is probably a combination of rapid radiation and substitution rate hererogeneity. Apparent incongruent patterns between different regions of the chloroplast genome are evaluated with ancient interspecific chloroplast recombination as explanatory model. Extremely elevated substitution rates in the exons of the plastid clpP gene was documented in Oenothera and three separate lineages of Sileneae. Introns have been lost in some of the lineages, but where present, intron sequences have a markedly slower substitution rate, similar to the rates found in other introns of their genomes. Three branches in the phylogeny show significant whole gene positive selection. In two of the lineages multiple partial copies of the gene were found.

<p>DNA sequences from the <i>rps16</i> intron and the <i>psbE-petL</i> spacer from the chloroplast genome, the ribosomal nuclear ITS region, and introns from the low copy nuclear genes <i>RPA2</i>, <i>RPB2</i>, <i>RPD2a</i> and <i>RPD2b</i>, are in different combinations used to infer phylogenetic relationships in <i>Sileneae</i> (<i>Caryophyllaceae</i>). Used in concert, the biparentally inherited nuclear regions are useful to distinguish between paralogy due to allopolyploidy and single gene duplications, respectively, because the latter are not expected to give rise to repeated phylogenetic patterns in potentially unlinked sequence regions. In addition, the sequences resolve previously poorly known relationships in the tribe <i>Sileneae</i>. Several independent losses and incomplete concerted evolution are inferred between the two <i>RPD2</i> paralogues in a subgroup of <i>Silene</i>.</p><p>An allopolyploid origin is suggested for the tetraploid <i>S. aegaea</i>, with the maternal ancestor from the diploid <i>S. pentelica</i> lineage, and the paternal contributor from the diploid <i>S. sedoides</i> lineage.</p><p><i>Silene involucrata</i> originated as an allotetraploid with the diploid lineage of Arctic <i>S. uralensis</i> as cytoplasmic donor and the diploid Siberian/Northeast Asian <i>S. ajanensis</i> lineage as pollen donor. A subsequent allopolyploidization with the <i>S. ajanensis</i> lineage as pollen donor and the tetraploid <i>S. involucrata</i> lineage as cytoplasmic donor resulted in the hexaploid lineage of <i>S. sorensenis sensu lato</i>.</p><p>A monophyletic origin of the North American polyploids is rejected. One lineage consists of tetraploid <i>S. menziesii</i> and its diploid allies. A separate lineage leads to a clade consisting of both diploid and polyploid Arctic, European and Asian taxa in addition to the majority of the North American polyploids. The tetraploid <i>S. californica</i> and the hexaploid <i>S. hookeri</i> are derived from separate allopolyploidization events between these two lineages.</p>

The focus for the first part of the thesis is on the systematics of species belonging to Silene subgenus Silene. Phylogenetic relationships are inferred from DNA sequences from both the plastid (the rps16 intron) and the nuclear (ITS, intron of the RPB2 gene) genomes. Silene section Rigidulae is shown to be non-monophyletic in its previous circumscription, but instead consisting of six separate clades, each correlated to the geographical distribution of the included species. The taxonomic consequences for each clade are discussed. One of the clades is recognized as a new section and described as Silene sect. Arenosae sect. nov. The morphological descriptions of the species are formalized using a novel implementation of the Prometheus Description Model. Two proposals are included in the thesis, one to reject the name Silene polyphylla L., which is a senior synonym to S. portensis L. Silene linearis Decne. is proposed for conservation against the rarely used S. linearis Sweet. Silene antirrhina, a weedy American annual, is strongly supported as sister to the Hawaiian endemic species of Silene, suggesting an American origin for these. Two of the endemics have evolved woodiness after introduction to Hawaii. In the second part of the thesis we use four nuclear DNA regions, (introns from RPA2, RPB2, RPD2a, RPD2b), and the chloroplast psbE-petG spacer. A framework is developed to evaluate different phylogenetic explanations for conflicting gene trees, where divergence times are used to discriminate among inter- and intralineage processes. The incongruences observed regarding the relationships among the three major lineages of Heliosperma are best explained by homoploid hybridization. The pattern regarding the origin of Heliosperma itself is more complicated and is likely to include several reticulate events. Two lineages have probably been involved in the origin of Heliosperma, one leading to Viscaria and Atocion and the other to Eudianthe and/or Petrocoptis.

DNA sequences from the rps16 intron and the psbE-petL spacer from the chloroplast genome, the ribosomal nuclear ITS region, and introns from the low copy nuclear genes RPA2, RPB2, RPD2a and RPD2b, are in different combinations used to infer phylogenetic relationships in Sileneae (Caryophyllaceae). Used in concert, the biparentally inherited nuclear regions are useful to distinguish between paralogy due to allopolyploidy and single gene duplications, respectively, because the latter are not expected to give rise to repeated phylogenetic patterns in potentially unlinked sequence regions. In addition, the sequences resolve previously poorly known relationships in the tribe Sileneae. Several independent losses and incomplete concerted evolution are inferred between the two RPD2 paralogues in a subgroup of Silene. An allopolyploid origin is suggested for the tetraploid S. aegaea, with the maternal ancestor from the diploid S. pentelica lineage, and the paternal contributor from the diploid S. sedoides lineage. Silene involucrata originated as an allotetraploid with the diploid lineage of Arctic S. uralensis as cytoplasmic donor and the diploid Siberian/Northeast Asian S. ajanensis lineage as pollen donor. A subsequent allopolyploidization with the S. ajanensis lineage as pollen donor and the tetraploid S. involucrata lineage as cytoplasmic donor resulted in the hexaploid lineage of S. sorensenis sensu lato. A monophyletic origin of the North American polyploids is rejected. One lineage consists of tetraploid S. menziesii and its diploid allies. A separate lineage leads to a clade consisting of both diploid and polyploid Arctic, European and Asian taxa in addition to the majority of the North American polyploids. The tetraploid S. californica and the hexaploid S. hookeri are derived from separate allopolyploidization events between these two lineages.

Heliosperma (nom. cons prop.) comprises 15—20 taxa, most of them endemic to the Balkan Peninsula. DNA sequences from the chloroplast (rps16 intron, psbE-petG spacer) and the nuclear genome (ITS and four putatively unlinked RNA polymerase genes) are used to elucidate phylogenetic relationships within Heliosperma, and its position within Sileneae. Three main lineages are found within Heliosperma: Heliosperma alpestre, H. macranthum and the H. pusillum-clade. The relationships among the lineages differ between the plastid and the nuclear trees. Relative dates are used to discriminate among inter- and intralineage processes causing such incongruences, and ancient homoploid hybridisation is the most likely explanation. The chloroplast data strongly support two, geographically correlated clades in the H. pusillum-group, whereas the relationships appear poorly resolved by the ITS data, when analysed under a phylogenetic tree model. However, a network analysis finds a geographic structuring similar to that in the chloroplast data. Ancient vicariant divergence followed by hybridisation events best explains the observed pattern. The morphological and taxonomical diversity in the H. pusillum-group is possibly ecology-induced, and is not correlated with the molecular data. Phylogenetic patterns regarding the origin of Heliosperma are complicated, probably influenced by reticulate and sorting events. At least two ancient lineages have been involved in its evolution, one most closely related to Viscaria/Atocion and the other to Eudianthe/Petrocoptis. Atocion and Viscaria are sister genera, most species-rich on the Balkans, and including six/three species. Phylogenies do not support their traditional classification, and provide a framework for a taxonomic revision. Atocion compactum is found in three different positions in the chloroplast tree, and in a single clade in the nuclear gene trees. Using relative dates we demonstrate that hybridisation with subsequent chloroplast capture is a feasible explanation for the pattern observed. This, and other observed reticulate patterns, highlights the importance of hybridisation in plant evolution.