Diet mixing enhances the performance of a generalist caterpillar,Platyprepia virginalis

1. Platyprepia virginalis caterpillars are dietary generalists and feed on multiple host species within a single day. We conducted field experiments to evaluate their performance on diets consisting of only their primary food, Lupinus arboreus, or diets consisting of L. arboreus plus other acceptable host species. 2. We found that relative growth rates and rates of survival were higher when they fed on mixed diets compared to lupine only. These results were consistent with hypotheses that mixed diets provided balanced nutrition, diluted toxins, and/or allowed recovery from parasitoids, although our data did not allow us to separate these non‐exclusive explanations. 3. We assayed alkaloids in their host foliage, in the caterpillars themselves, in parasitoids within caterpillars, in food boluses passing through their guts, and in frass that they excreted. We consistently found positive assays for alkaloids in foliage and in frass but negative assays in caterpillars, parasitoids, and food boluses. This suggests that the alkaloids that they ingest are metabolised or rendered non‐reactive by unknown means during passage through the gut. We found no support for the hypothesis that mixed diets prevented caterpillars from exhausting food supplies or allowed them to sequester chemicals from their alkaloid‐containing hosts. 4. Behavioural observations revealed that previous experience influenced a caterpillar's likelihood of moving to a different host. Caterpillars that had previously fed on other hosts were more likely to move to lupine while caterpillars previously collected on lupine were equally likely to choose more lupine or a different host. 5. These results are unusual in providing a clear and consistent benefit of diet mixing in a natural field setting where multiple ecological factors act upon the caterpillars.     

Behavioural ecology of defence in a risky environment: caterpillars versus ants in a Neotropical savanna

1. Predatory ants may reduce infestation by herbivorous insects, and slow‐moving Lepidopteran larvae are often vulnerable on foliage. We investigate whether caterpillars with morphological or behavioural defences have decreased risk of falling prey to ants, and if defence traits mediate host plant use in ant‐rich cerrado savanna. 2. Caterpillars were surveyed in four cerrado localities in southeast Brazil (70–460 km apart). The efficacy of caterpillar defensive traits against predation by two common ant species (Camponotus crassus, C. renggeri) was assessed through experimental trials using caterpillars of different species and captive ant colonies. 3. Although ant presence can reduce caterpillar infestation, the ants' predatory effects depend on caterpillar defence traits. Shelter construction and morphological defences can prevent ant attacks (primary defence), but once exposed or discovered by ants, caterpillars rely on their size and/or behaviour to survive (secondary defence). 4. Defence efficiency depends on ant identity: C. renggeri was more aggressive and lethal to caterpillars than C. crassus. Caterpillars without morphological defences or inside open shelters were found on plants with decreased ant numbers. No unsheltered caterpillar was found on plants with extrafloral nectaries (EFNs). Caterpillars using EFN‐bearing plants lived in closed shelters or presented morphological defences (hairs, spines), and were less frequently attacked by ants during trials. 5. The efficiency of defences against ants is thus crucial for caterpillar survival and determines host plant use by lepidopterans in cerrado. Our study highlights the effect of EFN‐mediated ant‐plant interactions on host plant use by insect herbivores, emphasizing the importance of a tritrophic viewpoint in risky environments.     

Adoption of lycaenid Niphanda fusca (Lepidoptera: Lycaenidae) caterpillars by the host ant Camponotus japonicus (Hymenoptera: Formicidae)

Caterpillars of the parasitic lycaenid butterfly are often adopted by host ants. It has been proposed that this adoption occurs because the caterpillars mimic the cuticular hydrocarbons of the host ant. This study aimed to examine whether caterpillars of the Japanese lycaenid butterfly Niphanda fusca induce adoption by mimicking their host ant Camponotus japonicus. Behavioral observations conducted in the laboratory showed that most second‐instar caterpillars were not adopted, whereas most third‐instar caterpillars were successfully adopted by host workers. A chemical comparison detected no characteristic differences in the cuticular hydrocarbon profiles between second‐ and third‐instar caterpillars. However, morphological features of the caterpillars differed between the second and third instars; third‐instar caterpillars developed exocrine glands (ant organs) such as tentacle organs and a dorsal nectary organ. These results suggest that multiple chemical signatures, not only cuticular hydrocarbons, may be important for invasion of the host ant nest.     

Effects of plant diversity and structural complexity on parasitoid behaviour in a field experiment

1. In natural ecosystems, plants containing hosts for parasitoids are often embedded within heterogeneous plant communities. These plant communities surrounding host‐infested plants may influence the host‐finding ability of parasitoids. 2. A release‐recapture‐approach was used to examine whether the diversity and structural complexity of the community surrounding a host‐infested plant influences the aggregation behaviour of the leaf‐miner parasitoid Dacnusa sibirica Telenga and naturally occurring local leaf‐miner parasitoids. Released and locally present parasitoids were collected on potted Jacobaea vulgaris Gaertn.plants infested with the generalist leaf‐miner Chromatomyia syngenesiae Hardy. The plants were placed in experimentally established plant communities differing in plant diversity (1–9 species) and habitat complexity (bare ground, mown vegetation, and tall vegetation). Additionally, parasitoids were reared out from host mines on the trap plants. 3. Plant diversity did not influence the mean number of recaptured D. sibirica or captures of other locally present parasitoids but the number of recaptured parasitoids was influenced by habitat complexity. No D. sibirica parasitoids were recaptured in the bare ground plots or plots with mown vegetation. The mean number of recaptured D. sibirica generally increased with increasing complexity of the plant community, whereas locally present parasitoids were captured more frequently in communities with more bare ground. There was a unimodal relationship between the number of reared out parasitoids and diversity of the surrounding vegetation with the highest density of emerged parasitoids at intermediate diversity levels. 4. The present study adds to the thus far limited body of literature examining the aggregation behaviour of parasitoids in the field and suggests that the preference of parasitoids to aggregate in complex versus simple vegetation is association specific and thus depends on the parasitoid species as well as the identity of the plant community.     

Host plant nitrogen enrichment has both positive and negative effects on the larval growth of a specialist butterfly

1. The nitrogen limitation hypothesis posits that phytophagous insects benefit from nitrogen enrichment of their host plants through a reduction of the concentration of toxic compounds and an increase of free amino acids and proteins. However, species' response to nitrogen enrichment varies substantially and high nitrogen levels are associated with population decline, suggesting there are major costs to feeding on nitrogen‐rich host plants. 2. To test the hypothesis that larval growth performance is maximal at intermediate nitrogen enrichment, nitrogen levels were measured in 18 populations of the host plant of Lycaena helle, a specialist butterfly inhabiting nutrient‐poor wet meadows. The nitrogen content of host plants was then modified to mirror average natural nitrogen levels (C), highest field‐recorded levels (T1), and levels higher than those observed across our study populations (T2). 3. Caterpillars fed with T1 leaves had a greater maximum body mass than caterpillars of the C group because of their improved food assimilation during the early stages of their development. Caterpillars of C and T2 groups had similar growth patterns but high nitrogen content had detrimental effects, as caterpillars fed with T2 leaves had a slower ingestion rate than C and T1 groups. 4. Quantifying the fitness consequences of these changes in growth performance is necessary to fully understand the implications of nitrogen enrichment for L. helle (rapid growth may result in fitness costs). However, conservation plans for this emblematic glacial relict species should also consider the preservation of its host plant quality to ensure its persistence.     

Associative learning of host presence in non‐host environments influences parasitoid foraging

1. Parasitoids are known to utilise learning of herbivore‐induced plant volatiles (HIPVs) when foraging for their herbivorous host. In natural situations these hosts share food plants with other, non‐suitable herbivores (non‐hosts). Simultaneous infestation of plants by hosts and non‐hosts has been found to result in induction of HIPVs that differ from host‐infested plants. Each non‐host herbivore may have different effects on HIPVs when sharing the food plant with hosts, and thus parasitoids may learn that plants with a specific non‐host herbivore also contain the host. 2. This study investigated the adaptive nature of learning by a foraging parasitoid that had acquired oviposition experience on a plant infested with both hosts and different non‐hosts in the laboratory and in semi‐field experiments. 3. In two‐choice preference tests, the parasitoid Cotesia glomerata shifted its preference towards HIPVs of a plant–host–non‐host complex previously associated with an oviposition experience. It could, indeed, learn that the presence of its host is associated with HIPVs induced by simultaneous feeding of its host Pieris brassicae and either the non‐host caterpillar Mamestra brassicae or the non‐host aphid Myzus persicae. However, the learned preference found in the laboratory did not translate into parasitisation preferences for hosts accompanying non‐host caterpillars or aphids in a semi‐field situation. 4. This paper discusses the importance of learning in parasitoid foraging, and debates why observed learned preferences for HIPVs in the laboratory may cancel out under some field experimental conditions.     

The mushroom habitat as an ecological arena for global exchange of Wolbachia

Wolbachia infect a variety of arthropod and nematode hosts, but in arthropods, host phylogenetic relationships are usually poor predictors of strain similarity. This suggests that new infections are often established by horizontal transmission. To gain insight into the factors affecting the probability of horizontal transmission among host species, we ask how host phylogeny, geographical distribution and ecology affect patterns of Wolbachia strain similarity. We used multilocus sequence typing (MLST) to characterize Wolbachia strain similarity among dipteran hosts associated with fleshy mushrooms. Wolbachia Supergroup A was more common than Supergroup B in Diptera, and also more common in mycophagous than non‐mycophagous Diptera. Within Supergroup A, host family within Diptera had no effect on strain similarity, and there was no tendency for Wolbachia strains from sympatric host species to be more similar to one another than to strains from hosts in different biogeographical realms. Supergroup A strains differed between mycophagous and non‐mycophagous Diptera more than expected by chance, suggesting that ecological associations can facilitate horizontal transmission of Wolbachia within mycophagous fly communities. For Supergroup B, there were no significant associations between strain similarity and host phylogeny, biogeography, or ecology. We identified only two cases in which closely related hosts carried closely related Wolbachia strains, evidence that Wolbachia‐host co‐speciation or early introgression can occur but may not be a major contributor to overall strain diversity. Our results suggest that horizontal transmission of Wolbachia can be influenced by host ecology, thus leading to partial restriction of Wolbachia strains or strain groups to particular guilds of insects.     

Non‐trophic effects of litter reduce ant predation and determine caterpillar survival and distribution

The tritrophic model featuring plants consumed by herbivores consumed by parasitoids or predators has become the primary paradigm used to describe herbivore dynamics. However, interactions involving herbivores can be habitat‐ specific and plants often provide habitat, as well as food. Structural complexity of the habitat may favor predators or may allow herbivore prey to escape detection and capture. This study considered the spatial and temporal dynamics of an arctiid caterpillar, Platyprepia virginalis. The tritrophic model that includes only a tachinid parasitoid that attacks P. virginalis and the caterpillars’ primary host‐plant, Lupinus arboreus, has failed to provide much insight into this system. Instead, we found that ants killed and removed many small caterpillars. Protecting caterpillars from ants increased their survival three‐fold and five‐fold in assays conducted during two years. Caterpillars were more likely to survive in short‐term assays at sites that naturally had a deeper cover of dead and living plant material. Experiments with baits showed that ant recruitment declined as litter depth increased on average. These survey results indicated that ant predation was an important source of mortality for young caterpillars and that the presence of thick litter reduced this mortality. These results were corroborated in an experiment that manipulated litter depth and ant access to caterpillars. Previous findings that other defoliating caterpillars increased litter depth and benefitted P. virginalis are also consistent with this hypothesis. Litter acts as an important non‐trophic resource, allowing caterpillars to avoid predation by ants such that wet sites with deep litter act as source populations for caterpillars. Our results show strong effects of both trophic and non‐trophic interactions since plants indirectly provided limiting habitat and this heterogeneous habitat strongly affected risk of predation and ultimately caterpillar abundance and distribution.     

Caterpillar assemblages on Chusquea bamboos in southern Ecuador: abundance,guild structure,and the influence of host plant quality

1. Information on the guild structure of foliage‐associated tropical insects is scarce, especially as caterpillars are mostly considered only as herbivores feeding on living leaves. However, many caterpillar species display alternative trophic associations, feeding on dead or withered leaves or epiphylls (‘non‐herbivores’). 2. To determine the contribution of these non‐herbivores, caterpillar communities associated with Chusquea Kunth (Poaceae) in the Andes of southern Ecuador were investigated. Caterpillars were collected at two elevation levels (montane rainforest ～2000 m and elfin forest at ～3000 m a.s.l.) and assigned to three feeding guilds (strict herbivores, non‐herbivores, and switchers) based on feeding trials. Foliage quality and leaf area were recorded to test for their influence on guild composition and caterpillar density. 3. Three hundred and eighty‐nine individuals belonging to 175 Lepidoptera species associated with Chusquea bamboos were found. The species richness of caterpillars was similarly high at both elevation levels but varied between feeding guilds. Approximately half (46.5%) displayed an alternative feeding association, i.e. were non‐herbivores (31.1%) or switchers (15.4%). 4. Caterpillar density was nearly two‐fold higher in the elfin forest, but only strict herbivores and switchers increased significantly with elevation. Leaf area positively influenced the density of strict herbivores and switchers; foliage quality only affected strict herbivores. The density of non‐herbivores did not differ significantly between the forest types and was not related to leaf area or foliage quality. 5. The present study underpins that non‐herbivores make up a considerable fraction of caterpillar communities in tropical mountain ecosystems and demonstrates that elevation, foliage quality and available plant biomass further shape feeding guild composition.     

Evolutionary history predicts high‐impact invasions by herbivorous insects

相似文献   

The effects of a multispecies sequential diet on the growth and survival of a tropical polyphagous caterpillar

Hylesia lineata Druce (Lepidoptera: Saturniidae) caterpillars are highly polyphagous at the individual level. I examined the effects of a multispecies sequential diet on the performance of the larvae. The experiment included three natural hosts respectively —Casearia corymbosa HBK,Thouinia paucidentata Radlk., andErythroxylum havanense Jacq — used by ovipositing females in proportion as expected by their abundance;C. corymbosa andE. havanense representing the most and least abundant hosts respectively. All possible two-host and single-host diets were tested. Larval survival to the pupal stage (as protected by a field cage) did not differ among all diets. Also, for females, regardless of the identity of the initial host, the final weight and the developmental time depended entirely on the nature of the final host. Male performance, in contrast, only differed in terms of developmental time but again the initial host did not affect the final result. It was concluded that a host switch early in the development of the caterpillars does not affect further growth on the other host plants. Caterpillars that finished their development onE. havanense reached small sizes but females devoted a larger proportion of their resources to egg biomass; as a result female fertility was similar among females regardless of their feeding history. I concluded that the caterpillar's abilities and the female compensatory flexibility may oppose selection for a strong preference hierarchy in the female ovipositing behavior.     

Host ontogeny determines parasitoid use of a forest caterpillar

For most organisms, patterns of natural enemy‐mediated mortality change over the course of development. Shifts in enemy pressure are particularly relevant for organisms that exhibit exponential growth during development, such as juvenile insects that increase their mass by several orders of magnitude. As one of the dominant groups of insect herbivores in most terrestrial plant communities, larval lepidopterans (caterpillars) are host to a diverse array of parasitoids. Previous research has described how the frequency of herbivore parasitism varies among host plants or habitats, but much less is known about how parasitism pressure changes during host development. To test whether the two major parasitoid taxa, wasps and flies, differentially attack shared hosts based on host developmental stage, we simultaneously exposed early‐ and late‐instar Euclea delphinii Boisduval (Lepidoptera: Limacodidae) caterpillars to parasitism in the field. We found strong evidence that parasitoids partition hosts by size; adult female wasps preferentially parasitized small caterpillars, whereas adult female flies preferred to attack large caterpillars. Our results demonstrate that host ontogeny is a major determinant of parasitoid host selection. Documenting how shifts in enemy pressure vary with development is important to understanding both the population biology and evolutionary ecology of prey species and their enemies.     

The origin and genetic differentiation of the socially parasitic aphid Tamalia inquilinus

Social and brood parasitisms are nonconsumptive forms of parasitism involving the exploitation of the colonies or nests of a host. Such parasites are often related to their hosts and may evolve in various ecological contexts, causing evolutionary constraints and opportunities for both parasites and their hosts. In extreme cases, patterns of diversification between social parasites and their hosts can be coupled, such that diversity of one is correlated with or even shapes the diversity of the other. Aphids in the genus Tamalia induce galls on North American manzanita (Arctostaphylos) and related shrubs (Arbutoideae) and are parasitized by nongalling social parasites or inquilines in the same genus. We used RNA sequencing to identify and generate new gene sequences for Tamalia and performed maximum‐likelihood, Bayesian and phylogeographic analyses to reconstruct the origins and patterns of diversity and host‐associated differentiation in the genus. Our results indicate that the Tamalia inquilines are monophyletic and closely related to their gall‐forming hosts on Arctostaphylos, supporting a previously proposed scenario for origins of these parasitic aphids. Unexpectedly, population structure and host‐plant‐associated differentiation were greater in the non‐gall‐inducing parasites than in their gall‐inducing hosts. RNA‐seq indicated contrasting patterns of gene expression between host aphids and parasites, and perhaps functional differences in host‐plant relationships. Our results suggest a mode of speciation in which host plants drive within‐guild diversification in insect hosts and their parasites. Shared host plants may be sufficient to promote the ecological diversification of a network of phytophagous insects and their parasites, as exemplified by Tamalia aphids.     

Companion planting – behaviour of the cabbage root fly on host plants and non-host plants

Six‐hundred individual female cabbage root flies (Delia radicum L.) (Diptera: Anthomyiidae) were each observed for 20 min under laboratory conditions to record how they behaved after landing on a host or a non‐host plant. Fly movements were recorded on host plants [cabbage –Brassica oleracea var. capitata (Cruciferae)] and non‐host plants [clover –Trifolium subterraneum L. (Papilionaceae)] surrounded by bare soil and on cabbage surrounded by clover. The most frequently observed behaviours made by the flies were (1) hops/spiral flights and (2) walks/runs. In the bare soil situation, the 50 individual flies observed in each treatment made 66 hops/spiral flights on the cabbage and 94 on the clover. When the two plants were tested together the movements were not additive as, instead of the expected 160 hops/spiral flights in the mixed plant treatment, the flies made 210 hops/spiral flights when they landed initially on cabbage but only 130 when they landed initially on clover. Few of the flies that landed initially on clover moved onto the host plant, even though the host plant was only a few centimetres away. The duration of the individual walks and runs made by the cabbage root flies were similar on both the host and non‐host plants. The only differences were the numbers of walks/runs made and the time the flies remained inactive. On the host plants, the females made four walks/runs, each of about 12 s duration, interspersed by rest periods that totalled 1.5 min. In contrast, on the non‐host plants the females made 10 walks/runs, each of about 9 s duration, interspersed by rest periods that totalled 7 min. Therefore, after landing on a plant, the flies, on average, left the host plant after 2.25 min and the non‐host plant after 8.5 min. Our conclusion is that the protracted time spent on the non‐host plants is the mechanism that disrupts insects from finding host plants in diverse plantings. Hence, the flies were arrested by non‐host plants rather than being repelled or deterred as suggested in earlier studies.     

Phytophagous insect-woody sprout interactions in tropical eucalypt forest. II. Insect community structure

Abstract Sucking insects constituted 79% of all phytophagous insects collected from woody sprouts in the ground layer of a tropical eucalypt forest. Mobile insect groups such as non-psyllid Hemiptera and Orthoptera were relatively frequent in this environment compared to temperate, Eucalyptus-dominated vegetation. The high fire frequency of the tropical eucalypt forest may favour mobile insect groups. The capture of sucking insects and caterpillars peaked in dry season samples. Other patterns of abundance of phytophagous insect groups showed little consistency in their seasonal trends between host species or between vegetation types within host species. Disparities between chewing insect abundance in daytime samples and the damage chewing insects cause, may result from disproportionate consumption by large, mainly nocturnal insects, such as members of the Orthoptera. In this study, 21% of insect species were specialists on single plant species. This study suggested that insect abundance reflected the growth patterns of woody sprouts after regular burning, rather than that plant growth and development were tuned to the pressures of insect herbivory.     

Use of habitat odour by host‐seeking insects

Locating suitable feeding or oviposition sites is essential for insect survival. Understanding how insects achieve this is crucial, not only for understanding the ecology and evolution of insect–host interactions, but also for the development of sustainable pest‐control strategies that exploit insects' host‐seeking behaviours. Volatile chemical cues are used by foraging insects to locate and recognise potential hosts but in nature these resources usually are patchily distributed, making chance encounters with host odour plumes rare over distances greater than tens of metres. The majority of studies on insect host‐seeking have focussed on short‐range orientation to easily detectable cues and it is only recently that we have begun to understand how insects overcome this challenge. Recent advances show that insects from a wide range of feeding guilds make use of ‘habitat cues’, volatile chemical cues released over a relatively large area that indicate a locale where more specific host cues are most likely to be found. Habitat cues differ from host cues in that they tend to be released in larger quantities, are more easily detectable over longer distances, and may lack specificity, yet provide an effective way for insects to maximise their chances of subsequently encountering specific host cues. This review brings together recent advances in this area, discussing key examples and similarities in strategies used by haematophagous insects, soil‐dwelling insects and insects that forage around plants. We also propose and provide evidence for a new theory that general and non‐host plant volatiles can be used by foraging herbivores to locate patches of vegetation at a distance in the absence of more specific host cues, explaining some of the many discrepancies between laboratory and field trials that attempt to make use of plant‐derived repellents for controlling insect pests.     

Varying degree of physiological integration among host instars and their endoparasitoid affects stress‐induced mortality

In natural populations of insect herbivores, genetic differentiation is likely to occur due to variation in host plant utilization and selection by the local community of organisms with which they interact. In parasitoids, engaging in intimate associations with their host during immature development, local variation may exist in host quality for parasitoid development. We compared the development of a gregarious endoparasitoid, Cotesia glomerata L. (Hymenoptera: Braconidae), collected in The Netherlands, in three strains and three caterpillar instars (L1–L3) of its main host, Pieris brassicae L. (Lepidoptera: Pieridae). Hosts had been collected in The Netherlands and France, and were reared in the laboratory for one generation. We also used an established Dutch laboratory strain that had not been exposed to parasitoids for at least 24 generations. Parasitoid survival to adulthood was inversely correlated with host instar at parasitism. Adult parasitoid body mass was largest when hosts were parasitized as L1 and smallest when hosts were parasitized as L3, whereas egg‐to‐adult development time was quickest on L3 hosts and slowest on L1 hosts. Higher survival and faster development of C. glomerata on French L2 hosts also showed that there is variation in host‐instar‐related suitability. Many L2 and most L3 caterpillars that were parasitized exhibited signs of pathogen infection and perished within a few days of parasitism, whereas this never happened when hosts were parasitized as L1 or in non‐parasitized control caterpillars. Our results reveal that, irrespective of the host strain, L1 hosts are optimally synchronized with C. glomerata development. By contrast, the high precocious mortality of L3 larvae may be due to stress‐induced regulation by the parasitoid in order to ‘force’ its developmental program into synchrony with the developing parasitoid larvae. Our results underscore a potentially important role played by pathogens in mediating herbivore–parasitoid interactions that are host‐instar‐dependent in their expression.     

Induced preference for host plant chemicals in the tobacco hornworm: contribution of olfaction and taste

Many herbivorous insects induce preferences for host plants. Recent work in Manduca sexta indicates that induced preferences are mediated by a “tuning” of the peripheral taste system to chemicals within host plant foliage. We tested this hypothesis by rearing caterpillars on artificial diet or potato foliage, and then examining olfactory- and taste-mediated responses to potato foliage extract. First, we confirmed earlier reports that consumption of potato foliage tunes the peripheral taste system by reducing responsiveness to glucose and increasing responsiveness to foliage extract. Second, we offered caterpillars a choice between disks treated with foliage extract (experimental) or solvent alone (control). The foliage-reared caterpillars approached and consumed the experimental disks disproportionately, whereas the diet-reared caterpillars approached and consumed both disks indiscriminately. This indicated that induced preferences involve olfaction and taste. Third, we ran choice tests with foliage-reared caterpillars deprived of either olfactory or gustatory input. Caterpillars lacking olfactory input approached both disks indiscriminately, but fed selectively on experimental disks. In contrast, caterpillars lacking gustatory input approached experimental disks selectively, but fed indiscriminately on both types of disk. We conclude that even though olfaction helps caterpillars locate potato foliage, it is the “tuned” gustatory response that ultimately mediates the induced preference.     

Host choice decisions in the polyembryonic wasp Copidosoma koehleri (Hymenoptera: Encyrtidae)

Female parasitoids often reject hosts of poor quality, where the survival and fitness of their offspring are expected to be low. In polyembryonic parasitoid wasps, a clone of genetically identical embryos develops from one egg in a host. In the wasp Copidosoma koehleri, each female clone produces one soldier larva that attacks competing clones inside the host. Aggression by soldiers is directed usually towards unrelated clones. Accordingly, it may be predicted that females will prefer nonparasitized over parasitized hosts, especially if the latter have been parasitized previously by a mated unrelated female, as a result of the reduced chances of survival for their offspring inside these hosts. In accordance with these predictions, females prefer nonparasitized hosts over self‐parasitized hosts when they are presented simultaneously. By contrast to the predictions, females prefer hosts parasitized by an unrelated conspecific over nonparasitized hosts when presented simultaneously. Females do not distinguish hosts parasitized by conspecifics from self‐parasitized hosts when presented simultaneously. They reject self‐parasitized hosts significantly more often than hosts parasitized by conspecifics when each host type is presented alone. Females faced with two previously parasitized hosts are not affected in their choice by the mating status (i.e. virgin or mated) of the previous parasitizing females. The combined results suggest that females are limited in their ability to assess the risk that their offspring will be attacked by a soldier, or that this risk is balanced by the relative advantages of ovipositing in a host parasitized by conspecifics. A possible advantage may be increased out‐breeding opportunities for the emerging offspring.