Search

In some parasitic Hymenoptera the dying caterpillars remain attached or close to the parasitoid cocoons. It has been suggested that the caterpillars act as ‘bodyguards’ for the vulnerable cocoons and therefore protect them against predators and/or hyperparasitoids (the ‘usurpation hypothesis’). This hypothesis has been demonstrated in associations where the caterpillars remain active and/or aggressive after parasitism. However, in other associations the caterpillars are so physiologically depleted after parasitism that they are unable to physically defend the cocoons and instead sit atop them in a moribund state. In this study a generalist predator, the spined soldier bug, Podisus maculiventris Say (Hemiptera: Pentatomidae), was provided with cocoons of the gregarious endoparasitoid Cotesia glomerata L. and the solitary endoparasitoid Microplitis mediator Haliday (both Hymenoptera: Braconidae), in turn attended by their hosts, Pieris brassicae L. (Lepidoptera: Pieridae) and Mamestra brassicae L. (Lepidoptera: Noctuidae), respectively. Cotesia glomerata produces broods of up to 40 cocoons and the dying caterpillars sit atop the cocoons where they exhibit little response to physical stimuli. Previous studies reported that dying P. brassicae caterpillars were ineffective bodyguards against two species of hyperparasitoids. In both associations, the dying host caterpillars were significantly preferred as food by P. maculiventris over the parasitoid cocoons. However, in absence of caterpillars, the bugs readily attacked the C. glomerata cocoons. Alternatively, the survival of M. mediator was very low, irrespective of whether a caterpillar was present or not. Caterpillars attacked by M. mediator are several times smaller than those attacked by C. glomerata. Consequently, the predators ran out of food much more quickly in the former and switched from one prey to the other. We show that in some host–parasitoid associations the dying caterpillars provide more visually apparent or nutritionally superior prey, rather than acting as bodyguards.

Summary Reproductive success for species in which offspring are confined to a distinct resource depends on the ability of parents to locate reproductive sites as well as the quality of these sites in terms of the food source, risk of predation and competition. To locate hosts for their offspring, parasitic wasps, or parasitoids, use plant odour blends induced by herbivore feeding. These herbivore-induced plant volatiles (HIPVs) may also be used by competitors and predators. Therefore, offspring of parasitoids that respond to the most conspicuous odours may find themselves more frequently involved in competition or predation risk. We studied cultivars of Brassica oleracea that are known to differ in volatile production that underlies attractiveness to parasitoids and asked whether variation in this parameter is associated with a heterogeneous distribution of intrinsic competition among parasitoid larvae and predation risk by hyperparasitoids that parasitize parasitoid larvae or pupae. We inoculated field-grown plants with Pieris caterpillars and, thereafter, exposed them to the natural parasitoid community. We measured the frequency of multiple incidences of parasitism in these herbivores. Cocoons of the parasitoids were collected to identify the degree of hyperparasitism associated with different Brassica cultivars. Pieris caterpillars on cultivars that were more attractive to Cotesia parasitoids were more commonly parasitized by several females of the same (superparasitism) or different wasp species (multiparasitism) than caterpillars on less attractive plants. Cocoons of parasitoids on attractive plants also more frequently produced hyperparasitoids. Our results show that there is heterogeneity in intrinsic competition and risk of hyperparasitism for parasitoids on different cabbage cultivars and that this heterogeneity is likely generated by variation in attraction of parasitoids to HIPVs of these cultivars. We conclude that parasitoids may find themselves between a rock and a hard place as cues for host presence may also predict high levels of competition and risk of predation. We speculate that this affects selection on parasitoid responses to plant odours and enhances selection on traits that make wasps better intrinsic or extrinsic competitors as well as selection for adaptive traits – such as crypsis – that protect them against hyperparasitoids.

The multicolored Asian lady beetle, Harmonia axyridis (Pallas) (Coleoptera: Coccinellidae), is an invasive species in North America that has a low susceptibility to indigenous natural enemies. We investigated the role of the immune system of H. axyridis in its resistance to North American population of the hymenopteran parasitoid Dinocampus coccinellae Schrank. An ultrastructural study of adult H. axyridis hemocytes under electron and light microscopy identified five different cellular types: plasmatocytes, granular hemocytes I, granular hemocytes II, oenocytoids, and spherule cells. Only granular hemocytes II and plasmatocytes are involved in the nodulation of bacteria and encapsulation of parasitoid eggs. Injection of Sephadex beads in H. axyridis adults induces an innate immune response and results in complete encapsulation of ≈28% of the beads after 24 h and 81% after 5 d. When H. axyridis was parasitized by D. coccinellae, we observed a decrease in the encapsulation response, and this pattern was inversely related to the number of parasitoid eggs in the H. axyridis hemolymph. The injection of polydnaviruses by D. coccinellae females and the multiple target hypotheses could both explain the encapsulation rate decreasing with increasing parasitoid egg load. Our results suggest that the immune system of H. axyridis could favor its escape from internal developing natural enemies indigenous to North America.

Co-evolutionary theory underpins our understanding of interactions in nature involving plant–herbivore and host–parasite interactions. However, many studies that are published in the empirical literature that have explored life history and development strategies between endoparasitoid wasps and their hosts are based on species that have no evolutionary history with one another. Here, we investigated novel associations involving two closely related solitary endoparasitoids that originate from Europe and North America and several of their natural and factitious hosts from both continents. The natural hosts of both species are also closely related, all being members of the same family. We compared development and survival of both parasitoids on the four host species and predicted that parasitoid performance is better on their own natural hosts. In contrast with this expectation, survival, adult size and development time of both parasitoids were similar on all (with one exception) hosts, irrespective as to their geographic origin. Our results show that phylogenetic affinity among the natural and factitious hosts plays an important role in their nutritional suitability for related parasitoids. Evolved traits in parasitoids, such as immune suppression and development, thus enable them to successfully develop in novel host species with which they have no evolutionary history. Our results show that host suitability for specialized organisms like endoparasitoids is closely linked with phylogenetic history and macro-evolution as well as local adaptation and micro-evolution. We argue that the importance of novel interactions and ‘ecological fitting’ based on phylogeny is a greatly underappreciated concept in many resource–consumer studies.

The performance of natural enemies, such as parasitoid wasps, is affected by differences in the quality of the host’s diet, frequently mediated by species or population-related differences in plant allelochemistry. Here, we compared survival, development time, and body mass in a generalist herbivore, the cabbage moth, Mamestra brassicae, and its solitary endoparasitoid, Microplitis mediator, when reared on two cultivated (CYR and STH) and three wild (KIM, OH, and WIN) populations of cabbage, Brassica oleracea. Plants either were undamaged or induced by feeding of larvae of the cabbage butterfly, Pieris rapae. Development and biomass of M. brassicae and Mi. mediator were similar on both cultivated and one wild cabbage population (KIM), intermediate on the OH population, and significantly lower on the WIN population. Moreover, development was prolonged and biomass was reduced on herbivore-induced plants. However, only the survival of parasitized hosts (and not that of healthy larvae) was affected by induction. Analysis of glucosinolates in leaves of the cabbages revealed higher levels in the wild populations than cultivars, with the highest concentrations in WIN plants. Multivariate statistics revealed a negative correlation between insect performance and total levels of glucosinolates (GS) and levels of 3-butenyl GS. However, GS chemistry could not explain the reduced performance on induced plants since only indole GS concentrations increased in response to herbivory, which did not affect insect performance based on multivariate statistics. This result suggests that, in addition to aliphatic GS, other non- GS chemicals are responsible for the decline in insect performance, and that these chemicals affect the parasitoid more strongly than the host. Remarkably, when developing on WIN plants, the survival of Mi. mediator to adult eclosion was much higher than in its host, M. brassicae. This may be due to the fact that hosts parasitized by Mi. mediator pass through fewer instars, and host growth is arrested when they are only a fraction of the size of healthy caterpillars. Certain aspects of the biology and life-history of the host and parasitoid may determine their response to chemical challenges imposed by the food plant.

Hyperparasitism by virgin female Encarsia tricolor was studied by direct observation of its behaviour when contacting two secondary host species (Encarsia formosa and E. tricolor) at different host stages (first and second larval stage, third larval stage, and pupal stage). The searching and hyperparasitism behavioural sequence of E. tricolor was independent of the host stage of the whitefly (Aleyrodes proletella), and was similar to several related primary parasitoid species. In experiments with equal numbers of secondary hosts, encounter frequencies were equal for both secondary host species in all developmental stages observed. However, rates of hyperparastism were different according to host stage and host species. Hosts in the late larval stages were most preferred for hyperparasitization and the heterospecific E. formosa was more preferred as a secondary host than the conspecific, E. tricolor, in particular from the prepupal stage onwards. The window of vulnerability, i.e., the duration of the period in which a secondary host is susceptible to hyperparasitism, was largely determined by the occurrence and rate of melanization after the onset of pupation. The duration of a successful hyperparasitization event was longer than one that failed. Superparasitism occurred only once in all cases. The potential effect of autoparasitoids on biological control programs and the consequences for selection and release of an effective, yet ecologically safe agent are discussed.

Insect parasitoids lay their eggs in arthropods. Some parasitoid species not only use their arthropod host for oviposition but also for feeding. Host feeding provides nutrients to the adult female parasitoid. However, in many species, host feeding destroys an opportunity to oviposit. For parasitoids that attack Homoptera, honeydew is a nutrient-rich alternative that can be directly imbibed from the host anus without injuring the host. A recent study showed that feeding on host-derived honeydew can be an advantageous alternative in terms of egg quantity and longevity. Here we explore the conditions under which destructive host feeding can provide an advantage over feeding on honeydew. For 5 days, Encarsia formosa Gahan (Hymenoptera: Aphelinidae) parasitoids were allowed daily up to 3 h to oviposit until host feeding was attempted. Host feedings were either prevented or allowed and parasitoids had ad libitum access to honeydew between foraging bouts. Even in the presence of honeydew, parasitoids allowed to host feed laid more eggs per hour of foraging per host-feeding attempt than parasitoids that were prevented from host feeding. The higher egg-laying rate was not compromised by survival or by change in egg volume over time. In conclusion, host feeding can provide an advantage over feeding on honeydew. This applies most likely under conditions of high host density or low extrinsic mortality of adult parasitoids, when alternative food sources cannot supply enough nutrients to prevent egg limitation. We discuss how to integrate ecological and physiological studies on host-feeding behavior.

Immature development of parasitoid wasps is restricted to resources found in a single host that is often similar in size to the adult parasitoid. When two or more parasitoids of the same or different species attack the same host, there is competition for monopolization of host resources. The success of intrinsic competition differs between parasitoids attacking growing hosts and parasitoids attacking paralyzed hosts. Furthermore, the evolution of gregarious development in parasitoids reflects differences in various developmental and behavioral traits, as these influence antagonistic encounters among immature parasitoids. Fitness-related costs (or benefits) of competition for the winning parasitoid reveal that time lags between successive attacks influence the outcome of competition. Physiological mechanisms used to exclude competitors include physical and biochemical factors that originate with the ovipositing female wasp or her progeny. In a broader multitrophic framework, indirect factors, such as plant quality, may affect parasitoids through effects on immunity and nutrition.

Constitutive and induced changes in plant quality impact higher trophic levels, such as the development 27 of parasitoids, in different ways. An efficient way to study how plant quality affects parasitoids is to 28 examine how the parasitoid larva is integrated within the host during the growth process. In two exper- 29 iments, we investigated the effects of varying nutritional quality of Brassica oleracea on parasitoid larval 30 development inside the host, the diamondback moth (Plutella xylostella). First, we compared larval 31 growth of the specialist Diadegma semiclausum and the generalist Diadegma fenestrale, when the host 32 was feeding on Brussels sprout plants that were either undamaged or were previously induced by cater- 33 pillar damage. Larvae of the generalist D. fenestrale showed lower growth rates than larvae of the special- 34 ist D. semiclausum, and this difference was more pronounced on herbivore-induced plants, suggesting 35 differences in host-use efficiency between parasitoid species. The growth of D. semiclausum larvae was 36 also analyzed in relation to herbivore induction on Brussels sprouts and on a wild B. oleracea strain. Par- 37 asitoid growth wasmore depressed on induced than on undamaged control plants, andmore on wild cab- 38 bage than on Brussels sprouts, which was largely explained by differences in host mass. The effects of 39 induction of wild Brassica on parasitoid development were pronounced early on, but as P. xylostella feed- 40 ing began inducing the previously undamaged control plants, the effect of induction disappeared, reveal- 41 ing a temporal component of plant–parasitoid interactions. This study demonstrates how insights into 42 the physiological aspects of host–parasitoid interactions can improve our understanding of the effects 43 of plant-related traits on parasitoid wasps.

Females of the larval parasitoid of Drosophila, Asobara citri, from sub-Saharan Africa, defend patches with hosts by fighting and chasing conspecific females upon encounter. Females of the closely related, palearctic species Asobara tabida do not defend patches and often search simultaneously in the same patch. The effect of patch defence by A. citri females on their distribution in a multi-patch environment was investigated, and their distributions were compared with those of A. tabida. For both species 20 females were released from two release-points in replicate experiments. Females of A. citri quickly reached a regular distribution across 16 patches, with a small variance/mean ratio per patch. Conversely, A. tabida females initially showed a clumped distribution, and after gradual dispersion, a more Poisson-like distribution across patches resulted (variance/mean ratio was closer to 1 and higher than for A. citri). The dispersion of A. tabida was most probably an effect of exploitation: these parasitoids increasingly made shorter visits to already exploited patches. We briefly discuss hypotheses on the adaptive significance of patch defence behaviour or its absence in the light of differences in the natural history of both parasitoid species, notably the spatial distribution of their hosts.

During their larval development, endoparasitoids are known to dispose of host resources in several different ways. Some parasitoid wasps consume most or all tissues of the host, whereas others consume a small fraction of host resources and either ensure that the host moves away from the pupation site or allow the host to remain close to the parasitoid cocoon(s). Using a single host species, Mythimna separata Walker (Lepidoptera: Noctuidae), this study compares the success of the two pupation strategies in the solitary parasitoids Microplitis sp. and Meteorus pulchricornis Wesmael (Hymenoptera: Braconidae) against attack from a secondary hyperparasitoid, Gelis agilis F. (Hymenoptera: Ichneumonidae). The caudal appendages of M. separata caterpillars parasitized by Microplitis sp. remain physically attached to parasitoid cocoons and the caterpillars behave aggressively when disturbed. However, after Me. pulchricornis larvae emerge from caterpillars of their host, M. separata, the parasitoid larvae pupate in cocoons that are suspended by a single thick thread that hangs 1–2 cm from under a leaf. In choice tests conducted in petri dishes, significantly fewer cocoons of Microplitis sp. attended by caterpillars than unattended cocoons were hyperparasitized by G. agilis. By contrast, Me. pulchricornis cocoons that were hanging from corn, Zea mays L., plants were hyperparasitized as frequently as those which were attached to leaves. We discuss the potentially different selection pressures generated among natural enemies such as predators and hyperparasitoids in determining optimal pupal defense strategies in primary parasitoids.

International audience; Interspecific competition amongst parasitoids is important in shaping the evolution of life-history strategies in these insects as well as community structure. Competition for hosts may occur between adult female parasitoids ('extrinsic' competition) or their progeny ('intrinsic' competition). Here, we examined intrinsic competition between two solitary secondary hyperparasitoids, Lysibia nana and Gelis agilis in cocoons of a primary parasitoid, Cotesia glomerata. Each species was allowed to sting hosts previously parasitized by the other at 24 h time intervals over the course of 144 h (6 days). When hosts were attacked simultaneously, neither species was dominant although the species to attack first won most encounters when it had a 24-48 h head start. However, after this time there was dramatic shift in the outcome with G. agilis dominating in all hosts > 72-h old, regardless of which species had parasitized C. glomerata first. G. agilis larvae, which initially had competed with L. nana for control of C. glomerata resources, began attacking the larvae of L. nana, whereas L. nana rejected hosts with older G. agilis larvae or pupae. Effects of multiparasitism also affected the development time and adult mass of the winning parasitoid. Our results reveal a shift in the trophic status of G. agilis from C. glomerata (in younger hosts) to L. nana (in older hosts), the first time such a phenomenon has been reported in parasitoids.

Larval development of the parasitoid Cardiochiles nigriceps Viereck occurs in the last instar larva of its host, Heliothis virescens (F.). This allows the parasitoid to exploit the nutritional increase in the biosynthetic activity occurring in the host in preparation for metamorphosis. To understand the biochemical basis of this host parasitoid developmental synchrony, we undertook host ligation studies and analyzed host hemolymph for proteins and glycerol esters. Parasitization affected the biochemical profile of the host. The hemolymph protein concentration of parasitized last instar H. virescens larvae increased through time, whereas unparasitized (control) larvae were characterized by a decrease in the protein titer when they reached the prepupal stage. The effect of parasitism on glyceride titers of host hemolymph was not as pronounced as the effect on proteins. Ligation conducted on 5th instar hosts, which were parasitized as 4th instars, affected parasitoid development in a time-dependent way. The percentage of successfully developing C. nigriceps larvae increased with the increase of the time interval between parasitization and ligation. Ligation performed before day 2 of the 5th larval instar of H. virescens completely inhibited parasitoid development. Ligations that disrupted parasitoid development were associated with a low host hemolymph protein concentration. Parasitoid development was successful when hemolymph protein titer was high, as occurred when ligations were performed after day 3 of the 5th host instar in both control and parasitized larvae. Ligations in both situations resulted in a slight increase in glyceride titers. The results suggest that host proteins and/or some factor(s) associated with them may play a role in parasitoid growth and development. (C) 1993 Wiley-Liss, Inc.

In natural systems, pre-adult stages of some insect herbivores are known to be attacked by several species of parasitoids. Under certain conditions, hosts may be simultaneously parasitised by more than one parasitoid species (= multiparasitism), even though only one parasitoid species can successfully develop in an individual host. Here, we compared development, survival, and intrinsic competitive interactions amongst three species of solitary larval endoparasitoids, Campoletis sonorensis (Cameron) (Hymenoptera: Ichneumonidae), Microplitis demolitor Wilkinson, and Microplitis croceipes (Cresson) (Hymenoptera: Braconidae), in singly parasitised and multiparasitised hosts. The three species differed in certain traits, such as in host usage strategies and adult body size. Campoletis sonorensis and M. demolitor survived equally well to eclosion in two host species that differed profoundly in size, Pseudoplusia includens (Walker) and the larger Heliothis virescens (Fabricius) (both Lepidoptera: Noctuidae). Egg-to-adult development time in C. sonorensis and M. demolitor also differed in the two hosts. Moreover, adult body mass in C. sonorensis (and not M. demolitor) was greater when developing in H. virescens larvae. We then monitored the outcome of competitive interactions in host larvae that were parasitised by one parasitoid species and subsequently multiparasitised by another species at various time intervals (0, 6, 24, and 48 h) after the initial parasitism. These experiments revealed that M. croceipes was generally a superior competitor to the other two species, whereas M. demolitor was the poorest competitor, with C. sonorensis being intermediate in this capacity. However, competition sometimes incurred fitness costs in M. croceipes and C. sonorensis, with longer development time and/or smaller adult mass observed in surviving wasps emerging from multiparasitised hosts. Our results suggest that rapid growth and large size relative to competitors of a similar age may be beneficial in aggressive intrinsic competition.

The larvae of most endoparasitoid wasps consume virtually all host tissues before pupation. However, in some clades, the parasitoid larvae primarily consume haemolymph and fat body and emerge through the side of the host, which remains alive and active for up to several days. The evolutionary significance of this host-usage strategy has attracted attention in recent years. Recent empirical studies suggest that the surviving larva guards the parasitoid broods against natural enemies such as predators and hyperparasitoids. Known as the 'usurpation hypothesis', the surviving larvae bite, regurgitate fluids from the gut, and thrash the head capsule when disturbed. In the present study, the 'usurpation hypothesis' is tested in the association involving Manduca sexta, its parasitoid Cotesia congregata, and a secondary hyperparasitoid Lysibia nana. Percentage parasitoid survival is higher and hyperparasitism lower when cocoons of C. congregata are attached to the dorsum of M. sexta caterpillars. Fat body contents in several associations involving solitary and gregarious parasitoids feeding on haemolymph and fat body are also compared. The amount of fat body retained in parasitized caterpillars varies considerably from one association to another. In M. sexta and Pieris brassicae, considerable amounts of fat body remain after parasitoid emergence whereas, in Cotesia kariyai and Cotesia rufricus, virtually all of the fat body is consumed by the parsasitoid larvae. The length of post-egression survival of parasitized caterpillars differs considerably in several tested associations. In Pseudeletia separata, most larvae die within a few hours of parasitoid emergence whereas, in M. sexta, parasitized larvae live up to 2 weeks after parasitoid emergence. Larvae in other associations parasitized by gregarious and solitary endoparasitoids live for intermediate periods. The results are discussed in relation to the adaptive significance of different feeding strategies of immature parasitoids and of the costs and benefits of retaining the parasitized caterpillar in close proximity with the parasitoid cocoons.

Caterpillars that are attacked by some species of parasitoid wasps are known to survive for several days after the parasitoid larvae emerge and pupate. It has been argued that the behaviour of the parasitized larva is ‘usurped’ by the parasitoid and that it ‘guards’ the parasitoid cocoons against their own natural enemies such as hyperparasitoids (the ‘usurpation hypothesis'). We tested this hypothesis in the association involving a gregarious endoparasitoid, the wasp Cotesia glomerata; caterpillars of its host, the large cabbage white butterfly Pieris brassicae; and a pupal hyperparasitoid, the wasp Lysibia nana. In laboratory experiments, we presented cocoon broods of C. glomerata to single females of L. nana in arenas for 6 h. We tested several treatments for rates of primary parasitoid survival, including variation in the position of the caterpillar and the presence or absence of an additional silk web spun by parasitized caterpillars. Parasitized P. brassicae larvae survived longer than the period necessary for C. glomerata adults to emerge. Rates of parasitoid survival were, however, unaffected by the presence of a P. brassicae larva on the cocoon brood, although significantly more parasitoids emerged when the silk web was present. Analyses of the foraging behaviour of individual L. nana females in arenas, performed using Observer software, revealed that the wasps showed a greater tendency to leave cocoons when caterpillars and silk were present. The laboratory experiments only partially support the usurpation hypothesis. In nature, usurpation of the host of the primary parasitoid may be a more effective strategy against generalist predators than against more specialized and better-adapted hyperparasitoids.

1. Changing plant composition in a community can have profound consequences for herbivore and parasitoid population dynamics. To understand such effects, studies are needed that unravel the underlying behavioural decisions determining the responses of parasitoids to complex habitats. 2. The searching behaviour of the parasitoid Diadegma semiclausum was followed in environments with different plant species composition. In the middle of these environments, two Brassica oleracea plants infested by the host Plutella xylostella were placed. The control set-up contained B. oleracea plants only. In the more complex set-ups, B. oleracea plants were interspersed by either Sinapis alba or Hordeum vulgare. 3. Parasitoids did not find the first host-infested plant with the same speed in the different environments. Sinapis alba plants were preferentially searched by parasitoids, resulting in fewer initial host encounters, possibly creating a dynamic enemy-free space for the host on adjacent B. oleracea plants. In set-ups with H. vulgare, also, fewer initial host encounters were found, but in this case plant structure was more likely than infochemicals to interfere with the searching behaviour of parasitoids. 4. On discovering a host-infested plant, parasitoids located the second host-infested plant with equal speed, demonstrating the effect of experience on time allocation. Further encounters with host-infested plants that had already been visited decreased residence times and increased the tendency to leave the environment. 5. Due to the intensive search of S. alba plants, hosts were encountered at lower rates here than in the other set-ups. However, because parasitoids left the set-up with S. alba last, the same number of hosts were encountered as in the other treatments. 6. Plant composition of a community influences the distribution of parasitoid attacks via its effects on arrival and leaving tendencies. Foraging experiences can reduce or increase the importance of enemy-free space for hosts on less attractive plants.

It is widely reported that plants emit volatile compounds when they are attacked by herbivorous insects, which may be used by parasitoids and predators to locate their host or prey. The study of herbivore-induced plant volatiles and their role in mediating interactions between plants, herbivores and their natural enemies have been primarily based on aboveground systems, generally ignoring the potential interactions between above and belowground infochemical- and food webs. This study examines whether herbivory by Delia radicum feeding on roots of Brassica nigra (black mustard) affects the behaviour of Cotesia glomerata, a parasitoid of the leaf herbivore Pieris brassicae, mediated by changes in plant volatiles. In a semi-field experiment with root-damaged and root-undamaged plants C. glomerata prefers to oviposit in hosts feeding on root-undamaged plants. In addition, in a flight-cage experiment the parasitoid also prefers to search for hosts on plants without root herbivores. Plants exposed to root herbivory were shown to emit a volatile blend characterized by high levels of specific sulphur volatile compounds, which are reported to be highly toxic for insects, combined with low levels of several compounds, i.e. beta-farnesene, reported to act as attractants for herbivorous and carnivorous insects. Our results provide evidence that the foraging behaviour of a parasitoid of an aboveground herbivore can be influenced by belowground herbivores through changes in the plant volatile blend. Such indirect interactions may have profound consequences for the evolution of host selection behaviour in parasitoids, and may play an important role in the structuring and functioning of communities.

A female parasitoid searching for hosts in a patch experiences a diminishing encounter rate with unparasitized and thus suitable hosts. To use the available time most efficiently, it constantly has to decide whether to stay in the patch and continue to search for hosts or to search for and travel to another patch in the habitat. Several informational cues can be used to optimize the searching success. Theoretically, encounters with unparasitized hosts should lead to a prolonged search in a given patch if hosts are distributed contagiously. The results of empirical studies strongly support this hypothesis. However, it has, to date, not been investigated theoretically whether encounters with already parasitized hosts (which usually entail time costs) provide a parasitoid with valuable information for the optimization of its search in depletable patches, although the empirical studies concerning this question so far have produced ambiguous results. Building on recent advances in Bayesian foraging strategies, we approached this problem by modeling a priori searching strategies (which differ in the amount of information considered) and then testing them in computer simulations. By comparing the strategies, we were able to determine whether and how encounters with already parasitized hosts can yield information that can be used to enhance a parasitoid's searching success.

1. Habitat complexity may stabilize interactions among species of different trophic levels by providing refuges to organisms of lower trophic levels. 2. Searching behaviour of the parasitoid, Diadegma semiclausum, was followed in different semifield set-ups, a low and high-density monoculture of Brassica oleracea and two intercrops, B. oleracea with Sinapis alba (also a member of the Brassicaceae) and B. oleracea with Hordeum vulgare (Poaceae). 3. When a low-density monocrop of B. oleracea was compared with a high-density monocrop, no differences were found in the ability of the female wasps to locate a host-infested plant, B. oleracea, infested with Plutella xylostella that was placed in the centre of the set-up. 4. The efficiency of the parasitoid to locate the host-infested plant was differentially affected by the species composition of the vegetation. Wasps entered the Sinapis-Brassica set-up faster, but took more time to find the host-infested plant than in the Hordeum-Brassica set-up. 5. The horizontal arrangement, i.e. by mixing S. alba or H. vulgare with, or placing them as rows between B. oleracea, did not affect host-finding efficiency. 6. Plant height did influence host finding. Wasps found the host-infested plants earlier in the set-up with short Sinapis plants compared with tall Sinapis plants. 7. Once the wasps had landed on the host-infested plant, the surrounding vegetation did not affect time needed to parasitize five consecutive hosts on the same infested plant, regardless of the composition or horizontal/vertical arrangement of the set-up. 8. Chemical and structural refuges in complex landscapes may play an important role in the persistence of this system through dampening oscillations of parasitoid and host populations. [KEYWORDS: crucifers ; Diadegma semiclausum ; Plutella xylostella ; proportional hazards model ; refuges]