Cynipid galls are atypical plant growths induced by wasps in the family Cynipidae that provide larvae with shelter and nutrition. Larvae gain control of attacked plant organs and send them on a new developmental trajectory, with three developmental phases known as initiation, growth, and maturation. Each of the approximately 1400 species of cynipid gall wasps manipulates plant tissues in a slightly different manner such that galls of each species are structurally distinct. Although the means by which cynipids initiate galls has fascinated naturalists for hundreds of years, the basic events in gall induction are still poorly understood. In an attempt to understand the galling strategies and developmental processes responsible for species-specific galls, I chose to compare the intimate details of life history strategies of two taxonomically related species attacking the same plant organs. The strategies included host specificity, phenology, and oviposition strategies, along with gall development to highlight basic events in gall biology and reveal possible stages in past speciation events. It is argued that differences in phenology and gall development played a role in driving speciation and thus, the patterns observed today are a result of extensive ecological interactions in the past that have influenced the evolution of these complex insect-plant relationships. Two species of cynipid wasps of the genus Diplolepis that occur on the wild roses of central Ontario proved to be ideal candidates for the study. One species, Diplolepis polita, induces single-chambered, prickly galls found in clusters on the adaxial surface of leaflets of Rosa acicularis whereas, the other species, D. nebulosa, induces single- iv chambered, smooth-surfaced galls found in clusters on the abaxial surface of the leaflets of Rosa blanda. Galls at all stages of development, from freshly oviposited eggs to maturation, were found in large numbers for both species. Leaf tissues from the bud stage to maturity of both species of rose were fixed in FAA, embedded in paraffin, sectioned and stained for histological study. Likewise, leaf tissues with freshly oviposited eggs of both species and galls from immaturity to maturity were fixed, sectioned, and stained. Comparing the several thousand slides made for the study revealed that galls of D. polita and D. nebulosa differ in their developmental events as well as the anatomy of their mature galls. Like the galls of all species of cynipids, those of D. polita and D. nebulosa are composed of distinct layers of gall cells known as nutritive, parenchymatous nutritive, sclerenchyma, cortex, and epidermis. Galls of D. polita consist of nutritive cells, parenchymatous nutritive cells, and an epidermis throughout the initiation and growth phases. Larvae remain small in relation to chamber volume until the maturation phase, when a hard layer of sclerenchyma differentiates. In contrast, galls of D. nebulosa have a delayed initiation phase, where galls remain nearly undetectable on leaflets for several weeks after oviposition before they enter the growth phase. Freshly-hatched larvae are protected by two layers of bowl-shaped patches of sclerenchyma that differentiates soon after initiation, along with nutritive cells and parenchymatous nutritive cells that surround the larval chambers, but galls of this species develop without an external layer of epidermis. Galls become spherical as they mature and a second layer of sclerenchyma differentiates within the walls of the galls, as does a layer of spongy cortex that appears between the second layer of sclerenchyma and the gall exterior. Larvae of D. nebulosa occupy nearly the entire volume of their larval chambers throughout gall development. v The size of each type of cell found within developing galls of both species, from gall initiation to gall maturity were measured and compared. Cells in the galls of both species continue to increase in size throughout development; however, the cells comprising the galls induced by D. polita are significantly larger than those induced by D. nebulosa. Comparing the biologies and galls of these two species, demonstrates how niche partitioning has occurred. It also shows that striking differences in structures occur within the galls of closely taxonomically related species as a result of differences in adult phenology, oviposition strategies, egg placement, and environmental conditions such as moisture levels. Although it has been proposed by other authors that diversity in cynipid galls results from differentiation of tissues found in the outer parts of galls, such as cortex and epidermis, the present study indicates the reasons are more complex. Furthermore, it is apparent that many aspects of gall development and anatomy have been overlooked by previous researchers and a variety of ecological factors contribute to differences in gall structure. To further complicate the already complex series of events that occur over the course of gall development, galls of D. polita and D. nebulosa are inhabited and structurally modified by inquilines of the genus Periclistus. These insects are also cynipid wasps, and have evolved a close relationship with Diplolepis galls whereby they kill the inducer larvae, feed on gall cells, and change the developmental trajectory of attacked galls. Galls of D. polita and D. nebulosa are attacked by two undescribed, but gall-specific species of Periclistus. Here, the inquiline associated with the galls of D. polita is referred to as Periclistus 1 and the inquiline associated with galls of D. nebulosa is referred to as Periclistus 2. The purpose of this study was to histologically examine all phases of vi modification by the two species of Periclistus to establish the events that are developmentally unique to inquilines. Periclistus 1 and 2 are phenologically distinct as Periclistus 1 oviposits into immature galls of D. polita in late May soon after galls are induced, and Periclistus 2 oviposits into immature galls of D. nebulosa in July. Modified galls of D. polita differ from modified galls of D. nebulosa as they are significantly enlarged compared to normal galls. Periclistus 1 chambers are arranged around the periphery of the inner gall whereas modified galls of D. nebulosa are of a similar size to normal galls. Chambers of Periclistus 2 are evenly distributed throughout the inner gall. Periclistus 1 and 2-modified galls undergo four phases of development identified as the egg phase, gall enlargement, chamber formation, and maturation phases. Both Periclistus 1 and 2 oviposit into immature galls, killing the inducer larvae with their ovipositors, and then the presence of Periclistus eggs along the inner chamber surface cause changes in gall structure. Diplolepis-induced nutritive cells degrade and Diplolepis-induced parenchymatous nutritive cells enlarge. Galls become significantly enlarged compared to those inhabited by inducer larvae and then feeding by first-instar Periclistus larvae stimulates the differentiation and proliferation of Periclistus-induced parenchymatous nutritive cells and nutritive cells. Immature larvae of both species of Periclistus initially feed around the inner surface of the Diplolepis-induced chamber, and then restrict their feeding to one spot. This results in cell proliferation such that each larva becomes restricted to the centre of a bowl-shaped growth of cells. Continued proliferation causes Periclistus nutritive and parenchymatous nutritive cells to rise up and completely encase the larvae. As this is occurring in modified galls of D. polita, a layer of sclerenchyma, referred to here as the vii inquiline-induced primary sclerenchyma, differentiates and circumscribes the periphery of the entire gall. This does not occur in galls of D. nebulosa until maturity. In modified galls of both species, nutritive cells and parenchymatous nutritive cells appear in dense clusters throughout the inside surface of Periclistus chambers. Once modified galls enter the maturation phase, inquiline-induced primary sclerenchyma differentiates, circumscribing the periphery of galls of D. nebulosa. In addition, Periclistus 1 and 2-inhabited galls both develop a second layer of inquiline-induced sclerenchyma, known as secondary sclerenchyma, around each inquiline chamber. Secondary sclerenchyma cells in the walls of Periclistus chambers are smaller than primary sclerenchyma cells circumscribing the entire gall. Gall cells induced by Periclistus 1 are larger than those induced by Periclistus 2; however, gall cells induced by both species of Periclistus are larger than those of their host Diplolepis galls. Based on differences in phenology, gall development, and final gall structure, modified galls of D. polita and D. nebulosa are anatomically distinct with each species of Periclistus responsible for gall tissues that are species-specific. In addition, the developmental pattern of Periclistus-modified galls is distinct from that of Diplolepis galls, illustrating the level of control inquilines have over the tissues of their host galls. This thesis demonstrates the complex nature of the interrelationships between cynipid wasps of the genera Diplolepis and Periclistus and their host roses. Diplolepis are true gall inducers that have an intimate relationship with the genus Rosa and there are many attributes of the genus Rosa that have contributed to the success of Diplolepis and allowed for their extensive radiation and divergence in their galls. Similarly, Periclistus viii inquilines have an intimate relationship with Diplolepis and the rose hosts. Periclistus have evolved the ability to manipulate rose tissues that have previously been under the influence of Diplolepis. Based on two species of inquilines examined in this thesis, Periclistus have also evolved an array of phenologies and modification strategies and like the developmental patterns of the Diplolepis, the developmental trajectories of Periclistus are also species-specific. This project has shown that the histological approach to studying cynipid galls, as well as those occupied by Periclistus inquilines, is highly rewarding and contributes to our overall understanding of these fascinating insects.

Cynipid gall formation is achieved by the intimate insect-plant interaction where by cynipid wasps redirect host plant development to form novel structures to protect and nourish the developing larva. To investigate the molecular mechanisms involved in this interaction, and extend our understanding of plant development, four approaches were taken. 1) A PCR based approach to search for genes to known signalling molecules: chitiooligosaccharides, or Nod factors, that control nodulation in the Rhizobia-legume interaction. PCR analysis was used to investigate the presence of the nodC gene in the cynipid gall wasp genome,h owever, no nodC-like sequencesw ere found. 2) SDS-PAGE analysis was carried out to compare inner-gall and non-gall protein signatures, demonstrating the variation between gall and non-gall tissue, and also that the protein signatures of inner-gall tissues vary between gall species. N-terminal sequencing and western blot analysis lead to the identification of a number of innergall proteins such as protein disulphide isomerase (PDI), formate dehydrogenase (FDH) and putative biotin carboxyl carrier protein (BCCP), involved in the synthesis of lipids in seeds. Analysis of the temporal and spatial expression of the putative BCCP revealed expression to be concentrated in the inner-gall cells throughout development, in all the gall species tested. 3) Cytological analysis of the inner-gall tissue was carried out throughout development of several gall species to investigate differences in their patterns of development and cytological characteristics of the inner-gall tissue, with many inner-gall cells being polytene. 4) A gall formation bioassay, to enable the activity of possible signals involved in gall formation to be tested, was developed. Rose callus tissue was used as a test tissue and the cynipid larval extract was exposed to this as a source of the active molecules. The induction of proteins in the callus after exposure to the larval extract was used as a molecular marker for activity. The polytene characteristic and the possible expression of seed proteins, suggest that seed developmental pathways may be used during gall formation.

iii Abstract Cynipid wasps of the genus Diplolepis Geoffrey induce galls exclusively on roses. These galls are susceptible to attack by various hymenopteran parasitoids, the most common of which are members of the family Eurytomidae, with 10 described species having been recorded in association with rose galls in Canada. Because of their small size, sexual dimorphism, morphological plasticity and poorly known biology, species identification of eurytomids is often uncertain. This thesis focused on identifying the eurytomids associated with the galls of the 14 native species of Diplolepis in Canada, testing species limits using a combination of morphological, mitochondrial DNA cytochrome c oxidase I (COI), ecological and geographical data. Six morphospecies were identified using morphological characters, while molecular data identified eight haplogroups. COI sequences identified one species of Tenuipetiolus Bugbee and seven species of Eurytoma Illiger, all associated with galls induced by Diplolepis in Canada. Inconsistencies were observed when compared to morphospecies, including four haplogroups within the genus Eurytoma that were overlooked by morphological studies. The morphospecies Eurytoma acuta Bugbee is a junior synonym of Eurytoma discordans Bugbee; however, the high intraspecific genetic divergences suggest the existence of a species complex. Two geographically sympatric but genetically distinct haplogroups were found within the morphospecies Eurytoma spongiosa Bugbee, “E. spongiosa 2” being a new species that was previously overlooked. An identification key along with (re)descriptions of all identified species was provided. This integrative taxonomy approach confirmed the oligophagous nature of most eurytomids, and provides new insight into the life history strategies of these parasitoids.

The British Isles have been invaded by 12 alien cynipid gallwasps over the past 150 years. The first 4 of these species have been studied in depth and represent a model system in phytophagous insect community structure. In this thesis, I extend this research programme to incorporate 8 further invaders. I examine recent changes in the distribution of invading oak gallwasps in Britain and spatial patterns in the composition of the associated communities of phytophagous cynipid inquilines and parasitoids. I use fully quantitative webs to assess the diversity and strength of trophic interactions between native and invading species and assess the potential for apparent competition between gallwasps mediated by shared natural enemies. Of the first 4 invaders to be studied, 3 have expanded their range since 1991/2. Three of these 4 species are now well established in Scotland, while Andricus corruptrix remains confined to England. Four new invaders (A. aries, A. lucidus, A. grossulariae, Aphelonyx cerricola) are established in southern England and are spreading. Rates of range expansion vary across species (between means of 3.3 and 24.4 km per year), and may be correlated with variation in lifecycles and abundance. The four newest invaders (Neuroterus saliens, Plagiotrochus australis, P. coriaceus, P. quercusilicis) are currently restricted to their sites of first record. Previous studies on one of the early invaders, Andricus quercuscalicis, identified south to north and east to west declines in community species richness and in the abundance of specific parasitoid species. I find that: 1) Parasitoid associations with the asexual galls of A. quercuscalicis track inquiline recruitment to this host. 2) The longitudinal and latitudinal gradients in parasitoid species richness demonstrated in previous work are no longer apparent, suggesting that younger northern communities may be converging on their older southern counterparts. 3) Inquilines show increasing survivorship with distance from the original centre of their distribution in south east England, suggesting at least temporary exploitation of enemy-free space. 4) The recently invading Andricus and Aphelonyx species have all rapidly recruited parasitoids and inquilines. Fully quantitative webs were constructed for 4 sites in England and Scotland incorporating both native and invading cynipids. I tested the hypothesis that newly arriving gallwasp generations would fall within food web compartments based on their host oaks and location on the tree as demonstrated in previous work. Counter to this hypothesis, parasitoids attacking one of the newest invaders (A. grossulariae) break down host tree-associated compartmentalisation. Where A. grossulariae has yet to become established, host-based compartmentalisation remains pronounced. Despite extensive sharing of parasitoid species, I found only one strong indirect interaction between species (both aliens) and no evidence for widespread apparent competition. Spatial density dependent predation on an appropriate scale can stabilise population dynamics. I quantified predation by blue tits (Parus caeruleus) of spring generation bud galls on Turkey oak (Q. cerris) at three spatial scales (shoots within branches, branches within trees, trees within sites). I found significant levels of bird predation, with most variation occurring between trees rather than between shoots within branches. Spatial density dependence was detected at sites in southern England, primarily at the level of trees within a site. Relationships at finer spatial scales were far more variable in magnitude and sign. My results suggest that blue tits forage primarily at the level of trees. This thesis presents comprehensive new data on the establishment and spread of 12 invading cynipid species and on their interactions with native communities. The results further understanding of both spatial and temporal aspects of natural enemy recruitment to invading species. In particular, it is clear that individual invading species can significantly modify trophic linkage between established food web compartments. Finally, my data emphasise the significant (but often unstudied) contribution of highly mobile vertebrate predators to otherwise closed ecological microcosms.