Many animal species are still undiscovered, even in Sweden. This thesis deals with the wasps of the subfamily Eucoilinae (of Figitidae, Cynipoidea), a poorly known group of small parasitoids of Dipteran flies. In this group, old classifications are chaotic, knowledge of old taxa is scant, phylogenetic work has recently started, and many new taxa remain to be described. In this thesis, an identification key is given to the European genera of Eucolinae, and a tribal classification of the subfamily is reintroduced, based on recent phylogenetic results. Most of the thesis focuses on the tribe Eucoilini. That tribe is analyzed phylogenetically in two analyses, one based on a large morphological dataset, the other on two molecular markers, ribosomal 28S and mitochondrial Cytochrome B. The monophyletic core group of Eucoilini is the Eucoila/Trybliographa complex. It is treated in a big taxonomic revision, based on studies of several hundred possibly relevant type specimens and several thousands of other specimens, and on the phylogenetic analyses. More than 500 names of possible members of this group are investigated, and many new combinations are proposed, of species found not to belong within the group, as well as those that did, including some which are transferred to other genera in accordance with the new phylogenetic results. The new genus Quasimodoana is erected, and 20 new species described (8 from Northern Europe, 12 from elsewhere). More than 180 new combinations, more than 50 new species-level synonyms and ca 10 genus-level synonyms, are proposed. Within this complex, the basal split is found to have been between a Paleotropic, specis-poor lineage (Bothrochacis), and a species-rich and largely Holarctic lineage, dominated by Trybliographa but also including the small genera Eucoila and Linoeucoila. The three are not satisfactorily resolved in the analysis, but still maintained as separate genera for pragmatic reasons. It is hypothesized that this group originated in Asia during the Eocene or Oligocene, attacking calyptrate flies in herbivore dung. Then it separated in a tropical branch retaining the ancestral life history mode, and a Holarctic branch, where some lineages specialised in Anthomyiidae and colonised the microhabitats of fungi and plants, speciating extensively in the boreal forests.

Arthropod species that have the potential to damage crops are food resources for communities of predators and parasitoids. From an agronomic perspective these species are pests and biocontrol agents respectively, and the relationships between them can be important determinants of crop yield and quality. The impact of biocontrol agents on pest populations may depend on the availability of other food resources in the agroecosystem. A scarcity of such resources may limit biological control and altering agroecosystem management to alleviate this limitation could contribute to pest management. This is a tactic of ‘conservation biological control’ and includes the provision of flowers for species that consume prey as larvae but require floral resources in their adult stage. The use of flowers for pest management requires an understanding of the interactions between the flowers, pests, biocontrol agents and non-target species. Without this, attempts to enhance biological control might be ineffective or detrimental. This thesis develops our understanding in two areas which have received relatively little attention: the role of flowers in biological control by true omnivores, and the implications of flower use by fourth-trophic-level life-history omnivores. The species studied were the lacewing Micromus tasmaniae and its parasitoid Anacharis zealandica. Buckwheat flowers Fagopyrum esculentum provided floral resources and aphids Acyrthosiphon pisum served as prey. Laboratory experiments with M. tasmaniae demonstrated that although prey were required for reproduction, providing flowers increased survival and oviposition when prey abundance was low. Flowers also decreased prey consumption by the adult lacewings. These experiments therefore revealed the potential for flowers to either enhance or disrupt biological control by M. tasmaniae. Adult M. tasmaniae were collected from a crop containing a strip of flowers. Analyses to determine the presence of prey and pollen in their digestive tracts suggested that predation was more frequent than foraging in flowers. It was concluded that the flower strip probably did not affect biological control by lacewings in that field, but flowers could be significant in other situations. The lifetime fecundity of A. zealandica was greatly increased by the presence of flowers in the laboratory. Providing flowers therefore has the potential to increase parasitism of M. tasmaniae and so disrupt biological control. A. zealandica was also studied in a crop containing a flower strip. Rubidium-marking was used to investigate nectar-feeding and dispersal from the flowers. In addition, the parasitoids’ sugar compositions were determined by HPLC and used to infer feeding histories. Although further work is required to develop the use of these techniques in this system, the results suggested that A. zealandica did not exploit the flower strip. The sugar profiles suggested that honeydew had been consumed by many of the parasitoids. A simulation model was developed to explore the dynamics of aphid, lacewing and parasitoid populations with and without flowers. This suggested that if M. tasmaniae and A. zealandica responded to flowers as in the laboratory, flowers would only have a small effect on biological control within a single period of a lucerne cutting cycle. When parasitoids were present, the direct beneficial effect of flowers on the lacewing population was outweighed by increased parasitism, reducing the potential for biological control in future crops. The results presented in this thesis exemplify the complex interactions that may occur as a consequence of providing floral resources in agroecosystems and re-affirm the need for agroecology to inform the development of sustainable pest management techniques.