It is probably a truism to say that what seems to be a single organism is, in fact, an assemblage of organisms – for there can hardly be an individual living outside a specialized laboratory that does not have commensals or parasites living within. The insects, the most diverse and numerous group of organisms on earth are no exception, and with micro- and macro-parasites from a wide range of taxa generate a remarkable range of interspecific associations. In some cases the insect is the sole host, in others it is an intermediate host or vector. It is the latter relationship which attracts much attention when insects and arachnids transmit infectious agents to humans, their animals or crops. Knowledge of the parasites of insects provides us with an opportunity to develop novel control methods for pests. Despite the diverse and widespread nature of insect infections, their impact on human well-being and the opportunity they give us to understand the complexity of the natural world, the subject remains a surprisingly neglected field. In this volume we endeavour to draw the veil from the 'black-box' approach to the insect stages of parasite life cycles to reveal some of the complexities of these relationships and how they are currently being analysed.

The literature contains many examples of changes induced by parasites in the behaviour and/or other phenotypic traits of insects. From an evolutionary perspective, the nature of these changes is usually difficult to assess. Parasite-induced changes in host behaviour can be adaptations of either host or parasite, or they can be mere pathological consequences of infection. Of the many criteria and experimental tests necessary to distinguish between adaptations and non-selected consequences, two are particularly important: the demonstration of fitness benefits for either host or parasite associated with the behavioural change, and the elucidation of the proximate mechanism responsible for the behavioural change. Another approach can serve to identify adaptive changes in behaviour: mapping specific behavioural alterations on a phylogeny of either hosts or parasites. The usefulness of this approach is illustrated with two examples, acanthocephalan- cockroach associations and insect-fungus associations. The adaptive nature of parasite-induced behavioural changes will always be difficult to evaluate because they are the product of two distinct but interacting genotypes. However, experimental and phylogenetic approaches can provide valuable insights into the evolutionary history of insect-parasite interactions.

Host fertility is often curtailed as a result of parasitic infection. The hypothesis that this may confer an adaptive advantage upon the symbionts if nutrients are directed from reproduction and made available for host/parasite maintenance is explored. The suggestion is made that an understanding of the mechanisms underlying the pathophysiology of fecundity reduction may shed light upon the evolutionary implications of this strategy for both parasite and host. To illustrate this the down-regulation of egg production is explored with reference to a particular model system, the association between metacestodes of the rat tapeworm, Hymenolepis diminuta and the mealworm beetle, Tenebrio molitor. Several aspects of host reproductive behaviour and physiology are affected by infection in this association, including vitellogenesis. Metacestodes directly inhibit the fat body synthesis of vitellogenin in a stage-specific, density-dependent manner. This inhibition is likely to be orchestrated by a modulator molecule, produced by the parasite. In the ovarian follicles, juvenile hormone III binding to a specific follicular membrane-binding protein is inhibited in infected beetles, resulting in the down-regulation of a cascade of events which enables vitellogenin to pass into the developing oocyte. Data to support the proposed existence of a parasite-induced antigonadotrophin, of host origin, are discussed. Evidence that similar mechanisms operate in Plasmodium-iniected anopheline mosquitoes and Onchocerca-infected blackflies is presented in support of the possibility that a parasite-induced reduction in host reproductive fitness is an adaptive strategy and an assessment of who is manipulating whom is made.

The effect of trypanosome infection on vector survival was observed in a line of Glossina morsitans morsitans selected for susceptibility to trypanosome infection. The differential effects of midgut and salivary gland infections on survival were examined by exposing flies to infection with either Trypanosoma congolense which colonizes midgut and mouthparts or Trypanosoma brucei rhodesiense which colonizes midgut and salivary glands. A comparison of the survival distributions of uninfected flies with those exposed to infection showed that salivary gland infection significantly reduces tsetse survival; midgut infection had little or no effect on the survival of tsetse. The significance of these findings is discussed in relation to the vectorial capacity of wild flies.

Host-parasitoid interactions are abundant in nature and offer great scope for the study of coevolution. A particularly fertile area is the interaction between internal feeding parasitoids and their hosts. Hosts have evolved a variety of means of combating parasitoids, in particular cellular encapsulation, while parasitoids have evolved a wide range of countermeasures. Studies of the evolution of host resistance and parasitoid virulence are reviewed, with an emphasis on work involving Drosophila and its parasitoids. Genetic variation in both traits has been demonstrated using isofemale line and artificial selection techniques. Recent studies have investigated the fitness costs of maintaining the ability to resist parasitoids, the comparative fitness of flies that have successfully defended themselves against parasitoids, and the degree to which resistance and virulence act against one or more species of host or parasitoid. A number of studies have examined geographical patterns, and sought to look for local adaptation; or have compared the traits across a range of species. Finally, the physiological and genetic basis of change in resistance and virulence is being investigated. While concentrating on Drosophila, the limited amount of work on different systems is reviewed, and other possible areas of coevolution in host-parasitoid interactions are briefly discussed.

Several recent studies have discussed the interaction of host life-history traits and parasite life cycles. It has been observed that the life-history of a host often changes after infection by a parasite. In some cases, changes of host life-history traits reduce the costs of parasitism and can be interpreted as a form of resistance against the parasite. In other cases, changes of host life-history traits increase the parasite's transmission and can be interpreted as manipulation by the parasite. Alternatively, changes of host's life-history traits can also induce responses in the parasite's life cycle traits. After a brief review of recent studies, we treat in more detail the interaction between the microsporidian parasite Edhazardia aedis and its host, the mosquito Aedes aegypti. We consider the interactions between the host's life-history and parasite's life cycle that help shape the evolutionary ecology of their relationship. In particular, these interactions determine whether the parasite is benign and transmits vertically or is virulent and transmits horizontally.

Key words: host-parasite interaction, life-history, life cycle, coevolution.

Parasitoids are parasites that invariably kill their host. Polydnaviruses are injected by parasitoid wasps into the body cavity of their insect host and cause immunosuppression, allowing the parasitoid to develop in the absence of encapsulation. One of the targets of the polydnaviruses are the haemocytes of the host, which undergo significant changes in response to entry of the virus. In some systems, haemocyte apoptosis is induced, or haemocyte clumping may be seen; in others, the cells round up and fail to adhere to a substrate. Effects on haemocytes may be transitory or permanent (cell death). Various polydnavirus gene products have been identified that interfere with normal haemocyte function. Phenoloxidase activity also is inhibited during parasitism, and the effect is inducible by polydnavirus. In some systems, venom components may act synergistically with polydnavirus in mediating the virally-induced effects on the host immune system. Polydnaviruses are powerful influences on the host immune system, which serve to permit successful development of the parasitoid without triggering the host immune response.

Reciprocal interactions of parasites transmitted by blood-sucking arthropod vectors have been studied primarily at the parasite–host and parasite–vector interface. The third component of this parasite triangle, the vector–host interface, has been largely ignored. Now there is growing realization that reciprocal interactions between arthropod vectors and their vertebrate hosts play a pivotal role in the survival of arthropod-borne viruses, bacteria, and protozoa. The vector–host interface is the site where the haematophagous arthropod feeds. To obtain a blood meal, the vector must overcome the host's inflammatory, haemostatic, and immune responses. This problem is greatest for ixodid ticks which may imbibe as much as 15 ml blood whilst continuously attached to their host for 10 days or more. To feed successfully, the interface between tick and host becomes a battle between the host's mechanisms for combating the tick and the tick's armoury of bioactive proteins and other chemicals which it secretes, via saliva, into the feeding lesion formed in the host's skin. Parasites entering this battlefield encounter a privileged site in their vertebrate host that has been profoundly modified by the pharmacological activities of their vector's saliva. For example, ticks suppress natural killer cells and interferons, both of which have potent antiviral activities. Not surprisingly, vector-borne parasites exploit the immunomodulated feeding site to promote their transmission and infection. Certain tick-borne viruses are so successful at this that they are transmitted from one infected tick, through the vertebrate host to a co-feeding uninfected tick, without a detectable viraemia (virus circulating in the host's blood), and with no untoward effect on the host. When such viruses do have an adverse effect on the host, they may impede their vectors' feeding. Thus important interactions between ticks and tick-borne parasites are displaced to the interface with their vertebrate host - the skin site of blood-feeding and infection.

Although there has been a great deal of research effort within the last two decades on identifying the active components of the saliva of blood-sucking ticks, mosquitoes, biting flies, fleas and bugs, essentially neglected have been the human lice. Despite initial reports in the early part of this century suggestive of vasodilatory, anticoagulant and immunosuppressive properties of the saliva, for the next 50 years there were no biochemical studies on the active principles. Very recently, anatomical and biochemical studies have begun to characterize the bioactive molecules in lice saliva. The louse stocks a salivary vasodilator in excess over what is needed for a single bite, and injects similar amounts at each successive bite. The vasodilator in lice saliva appears to have different pharmacological properties than peroxidative, oxidative and maxidilan types of vasodilators reported from other blood-sucking insects. Possible anticoagulant activities have also been characterized. This belated, but welcome, interest comes at a time of resurgence of lice-born disease in certain parts of Africa, and of resistance to chemical control in Europe and North America.

The ookinete is one of the most important stages of Plasmodium development in the mosquito. It is morphologically and biochemically distinct from the earlier sexual stages - gametocytes and zygote, and from the later stages - oocyst and sporozoites. Development to ookinete allows the parasite to escape from the tightly packed blood bolus, to cross the sturdy peritrophic matrix (PM), to be protected from the digestive environment of the midgut lumen, and to invade the gut epithelium. The success of each of these activities may depend on the degree of the biochemical and physical barriers in the mosquito (such as density of blood bolus, thickness of peritrophic matrix, proteolytic activities in the gut lumen etc.) and the ability of the ookinete to overcome these barriers. Ookinete motility, secretion of chitinase, resistance to the digestive enzymes, and recognition/invasion of the midgut epithelium all may play crucial roles in the transformation to oocyst. The overall sporogonic development of Plasmodium, therefore, depends on the results of the two-way manipulations between the parasite and the vector mosquito. Study of ookinete development and of the cellular and biochemical complexities of the mosquito gut may therefore lead to the design of novel strategies to block the transmission of malaria. This article reviews the intricate interactions between the parasite and the mosquito midgut in the context of development and transmission of Plasmodium parasites.

The most important function of any parasite is to secure transmission to new hosts. The gametocyte, the stage which has become developmentally committed to the sexual cycle, provides a critical link in the transmission of Plasmodium falciparum from the human host to the anopheline mosquito vector. It is therefore imperative that our determination to understand the biology of the gametocyte is greater than the technical obstacles which have resulted in the gametocyte being left very much out of the limelight by the intensive investigation of the asexual bloodstream parasite. Here we explore the areas of gametocyte biology which by nature of their relevance to control and pathology as well as basic biology, are the subjects of investigation in our laboratory. We also point out areas in need of particular attention.

Cytoplasmic incompatibility consists of sterility in cross matings, the crossing type being maternally inherited. It can be explained by the action of Wolbachia symbionts which are transmitted through the egg cytoplasm and leave an imprint on the sperm which prevents it fertilizing unless it is ‘rescued’ by the action of the same type of Wolbachia in the egg. Thus matings between infected males and uninfected females are sterile, but the reciprocal matings are fertile. Hence uninfected females are at risk of failing to transmit their uninfected cytoplasm if they cross mate, but infected females are at no such risk. Therefore natural selection favours the infected state and in two wild insect populations the infection has been observed spreading. If a gene for inability to transmit malaria could be introduced into Wolbachia and if this could be introduced into Anopheles (where these symbionts appear not to occur naturally), release of a limited number of such insects should trigger a process of displacement of malaria vectors, by the non-vector type. A simple model is used to demonstrate the limitations to this process which would be introduced by immigration.