Cynipid gall wasps have fascinating biology that has piqued the interest of naturalists throughout history. They induce morphologically complex, sometimes spectacular, gall structures on plants in which the larval stages develop. Gall wasps have therefore evolved an intimate association with their hosts - both metabolically, and in terms of their population histories. Gall wasps must both interact physiologically with their hosts to induce galls, and track their host plants through space and time. My thesis centres on two uses of genomic data in understanding the biology of the oak apple gall wasp Biorhiza pallida. I provide a comprehensive investigation into patterns of oak and gall wasp gene expression associated with gall induction, and a population genomic reconstruction of the population history of this species across the Western Palaearctic. While advances in sequencing technology and reduced costs have made these aims possible, analysis of the massive resulting datasets generated creates new challenges. Firstly, in reconstructing the population history of B. pallida, I describe the use of shotgun sequencing and an informatic pipeline to generate alignments of several thousand loci for three B. pallida individuals sampled from putative glacial refugia across the Western Palaearctic in Iberia, the Balkans and Iran. This dataset was analysed using a new maximum likelihood method capable of estimating population splitting and admixture among refugia across very large numbers of loci. The results showed an ancient divide between Iberia and the other two refugia, followed by very recent admixture between easternmost and westernmost regions. This suggests that gall wasps have migrated westwards along the North African coast as well as through mainland Europe. Second, I compare the gene expression profiles of gall wasp and oak tissues sampled from each of three stages of gall development, leading to new insights into potential mechanisms of gall wasp-oak interaction. A highly expressed gall wasp protein was identified that is hypothesised to stimulate somatic embryogenesis-like development of the gall through interaction with oak tissue glycoproteins. Highly expressed oak genes include those coding for nodulin-like proteins similar to those involved in legume nodule formation. Finally, analysis of the gall wasp genome has revealed potential, but as yet unconfirmed, horizontal gene transfer events into gall wasp genomes. Genes discovered in three gall wasp genomes and expressed in three transcriptomes encode plant cell wall degrading enzymes. They are not of hymenopteran origin, and are most homologous to genes of plant pathogenic bacteria. These genes could be involved in several aspects of gall wasp biology, including feeding and developmental manipulation of host plant tissue.

The interaction between figs (Ficus spp.) and their pollinating wasps (fig wasps; Chalcidoidea, Hymenoptera) is a classic example of an ancient and apparently stable mutualism. A striking property of this mutualism is that fig wasps consistently oviposit in the inner flowers of the fig syconium (gall flowers, which develop into galls that house developing larvae), but typically do not use the outer ring of flowers (seed flowers, which are pollinated and develop into seeds). This dissertation explores the potential differences between gall and seed flowers that might influence oviposition choices, and the unknown mechanisms underlying gall formation. To identify the microbial community that could influence oviposition choice, I identified fungi in both flower types across six species of Ficus. I found that whereas fungal communities differed significantly as a function of developmental stages of syconia and lineages of fig trees, communities did not differ significantly between receptive gall and seed flowers. Because secretions from the poison sac that are deposited at oviposition are thought to be important in gall formation by both pollinating fig wasps and non-pollinating, parasitic wasps, I examined poison sac morphology in diverse galling wasps from several species of Ficus in lowland Panama. I found that the size of the poison sac was positively associated with egg number across pollinating and non-pollinating fig wasps. Finally to determine difference in defense and metabolism between gall and seed flowers, and to identify genes involved in galling, I compared gene expression profiles of fig flowers at the time of oviposition choice and early gall development. I found a prominence of flavonoids and defensive genes in both pollinated and receptive gall flowers of Ficus obtusifolia, and revealed detectable differences between gall flowers and seed flowers before oviposition. Several highly expressed genes were also identified that have implications for the mechanism of gall initiation. This dissertation explores previously unstudied aspects of the fig and fig wasp mutualism and provides important molecular tools for future study of this iconic and ecologically important association.

One of the large lineages of parasitic wasps, the Cynipoidea, exhibits three distinctly different life modes. Slightly more than half of the about 3000 species are parasitoids in insect larvae, whereas the remaining species are associated with plants, either as gall inducers or as inquilines (guests feeding on plant tissue in galls). The main focus of this thesis has been to identify morphological changes associated with the shifts between life modes. Particular attention was paid to structures believed to be important in gall initiation. Comparative anatomical studies of the egg, larva and venom apparatus were performed, including representatives of parasitoids, gall inducers and inquilines. Examination of gross morphology and ultrastructure revealed that the eggs of the gall inducers are larger and surrounded by a thicker shell than the parasitoid eggs. These differences may be related to the fact that the gall inducer egg contains sufficient egg yolk for the embryo during the entire egg period, whereas the parasitoid egg often absorbs nutrients through the eggshell. Furthermore, the gall inducer egg is probably more exposed to desiccation and therefore a thicker and more resistant eggshell is crucial. Comparing the terminal-instar larvae of about 30 species of parasitoids, gall inducers and inquilines, extensive morphological variation was found, particularly in the head and mouthpart features. The variation was summarized in 33 morphological and one life-history character and parsimony analyses were performed. The resulting phylogenetic estimates were largely in accordance with previous analyses of adult morphology and molecular data. The larval data point to a single origin of the inquilines, in agreement with adult morphology but in conflict with molecular data. The venom apparatus was found to be quite uniform in structure among a sample of 25 species of cynipoid species. It consists of a very short venom duct, a reservoir and a single unbranched venom gland. With few exceptions, the venom apparatus is conspicuously larger relative to the female metasoma in the gall inhabiting species than in the parasitoids. We found little evidence of anatomical structures that could facilitate chemical communication between the gall-inducer embryo and the surrounding plant tissue through the thick eggshell. On the other hand, the enormous venom glands and reservoirs, which are apparently not used for defence, suggest that the adult female plays a significant role in gall induction by injecting secretions into the host plant when laying eggs.