A new cycad species, Cycas fushunensis sp. nov., is described from the Lutetian Jijuntun Formation at Fushun Coalmine, Liaoning Province, northeast China, based on a well-preserved partial frond containing about 15 leaflets. The fossil is characterized by a single strong vein per leaflet, decurrent leaflet base and haplocheilic stomata, suggesting that the fossil is attributed to the genus Cycas of Cycadaceae. Epidermal anatomical comparisons between the fossil and 17 selected modern Cycas species further indicate that C. fushunensis sp. nov. closely resembles Cycas panzhihuaensis Zhou et Yang, an endemic cycad to southwest China, due to characters shared, such as the straight anticlinal walls of both adaxial and abaxial epidermal cells and granular to striate cuticular characters on the internal surface of guard cell periclinal walls. The occurrence of close-to-modern Cycas from the early Cenozoic largely casts doubt on a hypothesis of the late Miocene differentiation of modern cycads, suggested by a recent molecular phylogenetic study.

The islands of the tropical South Pacific (TSP) are considered biodiversity hotspots. However, the biota of this region has received limited scientific attention and very little is known about its diversity, ecology and evolution. In this thesis we investigate some of the ecological and evolutionary processes in the TSP, focussing on lowland rain forests. We use molecular techniques to investigate evolutionary processes and vegetation surveys to study species diversity patterns and ecological processes. Chapter 1 reviews molecular, distributional and geographic evidence for dispersal versus vicariance explanations for the diversity and distribution of the TSP biota. Most islands of the TSP are geologically young (less than 40 million years old) and of oceanic origin, so most (if not all) of the biota on islands in the TSP arrived through long-distance dispersal events. This view is strongly corroborated by genetic data from published studies. Molecular studies also suggest two major source areas. One is located in the northwest, which includes Malesia and Southeast Asia, while the other is in the southwest and includes New Caledonia, Australia and New Zealand. We argue that local extinctions have occurred in source and stepping stone areas, creating sources of error for the interpretation of distribution and molecular data. In Chapter 2 we use allozyme data to investigate the question how Pacific cycads (Cycas, subsection Rumphiae) colonised the Pacific. We show that they colonised the Pacific and East Africa by long-distance dispersal, probably through floating seeds from a Malesian source area. Allozymes and morphological data provide support for two major groups within subsection Rumphiae and reveal close relationships between the extant species, suggesting very recent and/or ongoing dispersal events. Cycads are an example of recent diversification in a lineage with a long fossil record. The podocarp genus Dacrydium is another lineage with a long fossil record and in chapter 3 we investigate the colonisation and speciation processes in this lineage using allozymes and trnL-trnF plastid sequences. Our results suggest that the Pacific species of Dacrydium arrived recently (within the last 10 million years) in the TSP but are inconclusive about the source area of the genus. Combined molecular and ecological data suggest the occurrence of both allopatric and sympatric speciation in the Pacific radiation in this genus. Allozyme data also demonstrate the occurrence of hybridisation between two New Caledonian species. Our findings suggest that hybridisation and sympatric speciation may have played an important role in the evolution of the biota in the TSP. In chapter 4 we attempt to untangle the disparate forces driving alpha species diversity, forest structure and species composition in old-growth lowland tropical rainforest by assessing the tree species composition of twelve 1 ha vegetation count plots on 13 islands between New Guinea and Samoa. Using simplifications of a model based on biogeographic and ecological disturbance theory, we show that species diversity and richness are mainly influenced by size and area of an island, while endemism is mostly determined by isolation and area. High cyclone frequency is shown to increase the density of stems (with dbh > 10 cm). Correlations between the abundance of widespread canopy tree taxa and cyclone frequency suggest that cyclones affect species composition by increasing the abundance of cyclone-resistant species. However, floristic similarities show that geographic distance also affects species composition. It therefore appears that, for lowland rain forests in the TSP, biogeography is the major driver of species diversity and endemism and that disturbance is the major driver of forest structure, while both biogeography and disturbance affect the species composition. In chapter 5 we test the ability of NDVI (Normalised Difference Vegetation Index, a remotely sensed index of productivity) data and leaf samples as covariates of alpha species diversity using twelve vegetation count plots. NDVI performed poorly in estimating species diversity and species richness. However, the cost- and time-efficiency associated with remotely sensed data shows the potential of these methods, but only if accurate methods to estimate species richness are found. Species richness and species diversity estimates obtained from leaf litter samples correlate reasonably well with similar estimates obtained from count plots and are more than 30% cheaper and about 10% faster to obtain. If travel can be avoided through collaboration, leaf litter-based estimates of diversity could be obtained at about 5% the cost and in about half the time compared to count plots. Therefore the analysis of leaf litter is potentially a suitable and efficient method to obtain rapid estimates of species diversity in count plots. The final chapter discusses the roles of ecological and evolutionary processes in the TSP. While research to date has been scarce, especially on ecological processes acting on large scales, data show that the effects of ecology, evolution and biogeography are interlinked during the colonisation, establishment and subsequent evolution of taxa and biomes in the TSP.

Most gymnosperms produce a pollination drop that captures and transports pollen into the ovule. Pollination drops have other functions. These include influencing pollen germination and pollen tube growth, defending the ovule from pathogens and providing a food reward in insect-pollinated gymnosperms. Mineral and organic molecules, including proteins, are responsible for these additional functions. To date, pollination drops from a handful of conifers and one non-conifer gymnosperm, Welwitschia mirabilis, have been subjected to proteomic analysis. In the present study, tandem mass spectrometry was used to detect proteins in all gymnosperm lineages: cycads (Ceratozamia hildae, Cycas rumphii, Zamia furfuracea); Gnetales (Ephedra compacta, E. distachya, E. foeminea, E. likiangensis, E. minuta, E. monosperma, E. trifurca; Gnetum gnemon; Welwitschia mirabilis); Ginkgo biloba; conifers (Taxus x media). PEAKS 6 DB (Bioinformatics Solutions, Waterloo, ON, Canada) was used to make protein identifications. Proteins were detected in all gymnosperm species analyzed. The numbers of proteins identified varied between samples as follows: one protein in Welwitschia female; nine proteins in Cycas rumphii; 13 proteins on average in Ephedra spp.; 17 proteins in Gnetum gnemon; 38 proteins on average in Zamia furfuracea; 57 proteins in Ginkgo biloba; 61 proteins in Ceratozamia hildae; 63 in Taxus x media; 138 proteins in Welwitschia male. The types of proteins identified varied widely. Proteins involved in carbohydrate modification, e.g. galactosidase, chitinase, glycosyl hydrolase, glucosidase, were present in most gymnosperms. Similarly, defence proteins, e.g. reduction-oxidation proteins, lipid-transfer proteins and thaumatin-like proteins, were identified in many gymnosperms. Gymnosperms that develop a deep pollen chamber as the nucellus degrades, e.g., cycads, Ginkgo, Ephedra, generally contained higher proportions of proteins localized to intracellular spaces. These proteins represent the pollination drop degradome. Gymnosperms that either lack a pollen chamber, e.g. Taxus, or have a shallow pollen chamber, e.g. Gnetum, had greater proportions of extracellular proteins. These proteins represent the pollination drop secretome. Our proteomic analyses support the hypothesis that the pollination drops of all extant gymnosperms constitute complex reproductive secretions. / Graduate