Host plant exodus and larval wandering behaviour in a butterfly: diapause generation larvae wander for longer periods than do non‐diapause generation larvae

1. Prior to pupation, lepidopteran larvae enter a wandering phase lasting up to 30 h before choosing a pupation site. Because stillness is important for concealment, this behaviour calls for an adaptive explanation. 2. The explanation most likely relates to the need to find a suitable pupation substrate, especially in terms of shelter from predation, and given that many predators and parasitoids use host plants as prey‐location cues, mortality probably decreases with distance from the host plant. Hence, remaining on the host includes a long‐term risk, while moving away from the host introduces an increased risk during locomotion. 3. Bivoltine species that overwinter in the pupal stage produce two kinds of pupae; non‐diapausing pupae from which adults emerge after 1–2 weeks, or diapausing pupae that overwinter with adults emerging after 8–10 months. 4. Given the hypothesis of distance‐from‐host‐plant‐related predation, this should select for phenotypic plasticity with larvae in the diapausing generation having a longer wandering phase than larvae under direct development, if there is a trade‐off between mortality during the wandering phase and accumulated mortality during winter. 5. Here this prediction is tested by studying the duration of the wandering period in larvae of the partially bivoltine swallowtail butterfly, Papilio machaon, under both developmental pathways. 6. The results are in agreement with the predictions and show that the larval wandering phase is approximately twice as long under diapause development. The authors suggest that the longer duration of the wandering phase in the diapause generation is a general phenomenon in Lepidoptera.     

Spatial model for predicting the presence of cinnamon fungus (Phytophthora cinnamomi) in sclerophyll vegetation communities in south‐eastern Australia

Abstract The pathogen Phytophthora cinnamomi causes extensive ‘dieback’ of Australian native vegetation. This study investigated the distribution of infection in an area of significant sclerophyll vegetation in Australia. It aimed to determine the relationship of infection to site variables and to develop a predictive model of infection. Site variables recorded at 50 study sites included aspect, slope, altitude, proximity to road and road characteristics, soil profile characteristics and vegetation attributes. Soil and plant tissues were assayed for the presence of the pathogen. A geographical information systyem (GIS) was employed to provide accurate estimations of spatial variables and develop a predictive model for the distribution of P. cinnamomi. The pathogen was isolated from 76% of the study sites. Of the 17 site variables initially investigated during the study a logistic regression model identified only two, elevation and sun‐index, as significant in determining the probability of infection. The presence of P. cinnamomi infection was negatively associated with elevation and positively associated with sun‐index. The model predicted that up to 74% of the study area (11 875 ha) had a high probability of being affected by P. cinnamomi. However, the present areas of infection were small, providing an opportunity for management to minimize spread into highly susceptible uninvaded areas.     

Distribution,prevalence and mean numbers of larvalAnisakis simplex (Nematoda: Ascaridoidea) in Atlantic salmon,Salmo salar L. and their use as biological indicators of host stocks

Synopsis An estimate of the arithmetic mean number of larvalAnisakis simplex per salmon,Salmo salar, in the north Atlantic was obtained from data collected off west Greenland where salmon from both sides of the Atlantic intermingle. An average of 5.34 larvae was found in the body cavity or associated viscera of 771 age 1 salmon representing some 60.8% of the total number of larvae present. LarvalA. simplex were found at all sampling stations throughout the north Atlantic. In most samples prevalence varied only slightly from area to area.The mean numbers of larvae in the body cavity and on associated organs of salmon in samples from North America were significantly different between samples. Comparisons of the mean numbers of larvae with several known variables in the data were, therefore, considered: host's sex, host's sea-age, geographic locality, year and season of capture.There was evidence that the mean number of larvae per host was greater in salmon from Europe than in those from North America, but the data could not be used to providereliable estimates of the mean numbers of larvae from each area because of heterogeneity within areas. Thus the data from west Greenland and from continental home waters could net be used in calculations to determine the percentage composition of North American and European salmon in the Greenland fishery although they supported earlier findings of mixing of salmon stocks off west Greenland.The use of parasitological data in the identification of different Atlantic salmon stocks is discussed.     

Environmental factors associated with the distribution of Anopheles arabiensis and Culex quinquefasciatus in a rice agro-ecosystem in Mwea, Kenya

Studies were conducted between May and June, 2006 to investigate the environmental factors affecting the distribution of An. arabiensis Patton and Culex quinquefasciatus Say in Mwea, Kenya. The sampling unit comprised all non-paddy aquatic habitats and ten randomly selected paddies and canals located within a 200 m radius from the periphery of the study site. Thirteen physico-chemical variables were recorded for each sampling site in each sampling occasion and a sample of mosquito larvae and other aquatic invertebrates collected. The non-paddy aquatic habitats identified included pools and marshes. Morphological identification of 1,974 mosquito larvae yielded four species dominated by Cx. quinquefasciatus (73.2%) and An. arabiensis (25.0%). Pools were associated with significantly higher Cx. quinquefasciatus larval abundance and less diversity of other aquatic invertebrates compared with other habitat types. In contrast, the abundance of An. arabiensis did not differ significantly among habitat types. Culex quinquefasciatus habitats had higher water conductivity and exhibited a higher abundance of other aquatic invertebrates than An. arabiensis habitats. Chi-square analysis indicated that the two species were more likely to coexist in the same habitats than would be expected by chance alone. Anopheles arabiensis larvae were positively associated with dissolved oxygen and adults of family Haliplidae and negatively associated with emergent vegetation and Heptageniidae larvae. Culex quinquefasciatus larvae were positively associated with dissolved oxygen, total dissolved solids, Chironomidae larvae, and Microvelidae adults and negatively associated with emergent vegetation. These findings suggest that both biotic and abiotic factors play a significant role in niche partitioning among Cx. quinquefasciatus and An. arabiensis, a factor that should be considered when designing an integrated vector control program.     

Substrate selection by saprophagous wood‐borer larvae within highly variable hosts

In insects completing their larval development within a single host, oviposition site is seen as a major determinant of offspring performance. However, in previous studies, the saprophagous wood‐borer Anthophylax attenuatus (Haldeman) (Coleoptera: Cerambycidae) showed no strong response to between‐host variations in nutritional factors influencing larval growth and survival. To explain such weak selection in adults, we hypothesized that substrate selection occurs at a smaller scale by larvae within hosts showing high variability in substrate quality. In this study, we described within‐host variability in wood density and determined whether wood‐boring larvae were found more often than expected in specific decay types. We characterized the variability of decay in 24 snags by producing wood density profiles for each. We then collected larvae from the same snags through wood dissection, and associated a wood density value to each by taking a wood sample around each larva found. We then compared ratios of available and used substrate types defined by wood density. We observed substantial within‐snag variation in wood density. Middle decay class (0.275–0.375 g cm^?3) was significantly overused by larvae, whereas more decayed wood was clearly avoided. High within‐host variability in substrate quality and active or passive selection by larvae of specific substrate types suggest that selection pressures on adult behaviour could be lower than expected for a parasitic species, and might be linked with the weak selection observed at a larger scale by ovipositing adults.     

Selection for enemy‐free space: eggs placed away from the host plant increase survival of a neotropical ithomiine butterfly

1. The selection of an oviposition site by a phytophagous insect can depend on many factors, including the risk of predation. Many species avoid predators by laying eggs where enemies searching host plants are unlikely to find them. 2. Females of the Peruvian butterfly, Oleria onega Hewitson (Lepidoptera: Nymphalidae: Danainae: Ithomiini) lay most of their eggs (76 ± 9%) off the host plant, Solanum mite Ruiz & Pav. These off‐host eggs may be laid up to 0.5 m from the nearest host‐plant individual, on twigs or leaf litter, as well as on living plants of species unsuitable for larval food. 3. Disappearance of eggs on and off the host plant was recorded by transferring eggs laid in captivity to known locations in the wild and recording rates of disappearance before the larvae emerged. After 2 days, eggs on the host were significantly more likely to have disappeared compared to eggs laid elsewhere. 4. We conclude that a high risk of predation is a likely trigger that caused O. onega to evolve a behaviour of laying eggs off its host plant.     

A comparison of the host preference of monarch butterflies (Danaus plexippus) for milkweed (Asclepias syriaca) over dog-strangler vine (Vincetoxicum rossicum)

Observations in the field indicate that monarch butterflies will oviposit on dog‐strangler vine, an invasive introduced species in the same family as milkweed (Asclepias spp.), the principal larval host of monarchs. The potential impact of this behaviour depends on the strength of the preference of monarch adults to oviposit on these two hosts and the relative ability of larvae to survive on each. We determined the preference for milkweed vs. dog‐strangler vine of ovipositing adults and first instar larvae in choice and no‐choice tests. We also compared the ability of larvae to consume, develop, and survive on either host. In the presence of both hosts, adults exhibited a strong preference to oviposit on milkweed over dog‐strangler vine (mean 80.7 eggs compared to 0.4 eggs over 48 h, respectively). In the absence of milkweed, adults ceased oviposition (mean 0.9 eggs in 48 h), but resumed oviposition when the dog‐strangler vine was replaced with milkweed (mean 99.1 eggs in 48 h). Given a choice between hosts over 24 h, 92% of larvae moved to milkweed leaves and consumed 3.94 cm² of milkweed leaves compared to 2% of larvae that moved to dog‐strangler vine and consumed negligible amounts of leaf material (0.01 cm²). Without a choice, larvae on dog‐strangler vine never consumed more than mean 0.02 cm² larva^?1 in a 24‐h period, did not develop beyond the first instar, and died within 96 h. We obtained no data in support of an effect of the presence of dog‐strangler vine on monarch butterfly populations.     

Seasonal abundance and activity of the tick Dermacentor silvarum in northern China

Studies on the seasonal abundance and activity of larvae, nymphs and adults of Dermacentor silvarum (Ixodida: Ixodidae), both on and off the host, were carried out over a 2-year period from February 2008 to January 2010 in northern China. During the study period, feeding ticks were removed weekly from domestic sheep and free-living ticks were collected weekly by flag-dragging in four different habitat types. Rodents were captured and inspected for immature ticks from June to September 2008. The results of this study suggest that the ticks mainly reside in shrubs and complete only one generation per year with sequential seasonal distribution and little overlap between the activity of the larvae and nymphs. Adults were most prevalent from late February to late May and peaked in number in mid-April. Larvae were found from early June to early September and peaked in number in mid-July. Nymphs were mainly distributed from late June to late September and peaked in mid-August. Adult and nymphal D. silvarum were found primarily on the ears of sheep, but no larvae were found on sheep. Additionally, an overwintering male adult population was detected on sheep after October, but no free-living adults were found by dragging.     

D-vac sampling for predatory arthropods in winter wheat

We evaluated the D-vac suction machine for sampling predatory arthropods in Oklahoma winter wheat fields. The efficiency of D-vac sampling was low for adult Coccinellidae and Carabidae. Sampling efficiency was greater for coccinellid larvae. Sampling efficiency was high for adult and immature Nabidae and Chrysopidae, for Araneae, and for adult Staphylinidae. For most predators, there were significant correlations between the number of individuals in D-vac samples and the number of individuals per m² in the field. The highest correlation was 0.82 for adult Staphylinidae. Correlations for most predators were <0.60. Significant partial correlations between D-vac sample estimates and ancillary abiotic and biotic variables occurred for most predators, indicating that D-vac sampling efficiency was affected environmental variables. Multiple regression models were constructed to relate population estimates from D-vac sampling to absolute density by adjusting estimates for influential abiotic and biotic environmental variables. Significant regression models were not achieved for adult Coccinellidae or adult Carabidae, and the coefficient of determination was low (0.18) for the model for adult Nabidae. Higher values of R² were achieved for larval Coccinellidae, Chrysopidae adults and larvae, Nabidae nymphs, Araneae, and adult Staphylinidae. Among abiotic variables, air temperature and wind velocity most frequently entered into step-wise regression models. Among biotic variables, wheat plant growth stage and wheat tiller density frequently entered into models.     

Selecting discriminant function models for predicting the expected richness of aquatic macroinvertebrates

1. The predictive modelling approach to bioassessment estimates the macroinvertebrate assemblage expected at a stream site if it were in a minimally disturbed reference condition. The difference between expected and observed assemblages then measures the departure of the site from reference condition. 2. Most predictive models employ site classification, followed by discriminant function (DF) modelling, to predict the expected assemblage from a suite of environmental variables. Stepwise DF analysis is normally used to choose a single subset of DF predictor variables with a high accuracy for classifying sites. An alternative is to screen all possible combinations of predictor variables, in order to identify several ‘best’ subsets that yield good overall performance of the predictive model. 3. We applied best‐subsets DF analysis to assemblage and environmental data from 199 reference sites in Oregon, U.S.A. Two sets of 66 best DF models containing between one and 14 predictor variables (that is, having model orders from one to 14) were developed, for five‐group and 11‐group site classifications. 4. Resubstitution classification accuracy of the DF models increased consistently with model order, but cross‐validated classification accuracy did not improve beyond seventh or eighth‐order models, suggesting that the larger models were overfitted. 5. Overall predictive model performance at model training sites, measured by the root‐mean‐squared error of the observed/expected species richness ratio, also improved steadily with DF model order. But high‐order DF models usually performed poorly at an independent set of validation sites, another sign of model overfitting. 6. Models selected by stepwise DF analysis showed evidence of overfitting and were outperformed by several of the best‐subsets models. 7. The group separation strength of a DF model, as measured by Wilks’Λ, was more strongly correlated with overall predictive model performance at training sites than was DF classification accuracy. 8. Our results suggest improved strategies for developing reliable, parsimonious predictive models. We emphasise the value of independent validation data for obtaining a realistic picture of model performance. We also recommend assessing not just one or two, but several, candidate models based on their overall performance as well as the performance of their DF component. 9. We provide links to our free software for stepwise and best‐subsets DF analysis.     

Unexpected collective larval dispersal but little support for sweepstakes reproductive success in the highly dispersive brooding mollusc Crepidula fornicata

In many marine invertebrates, long‐distance dispersal is achieved during an extended pelagic larval phase. Although such dispersal should result in high gene flow over broad spatial scales, fine‐scale genetic structure has often been reported, a pattern attributed to interfamilial variance in reproductive success and limited homogenization during dispersal. To examine this hypothesis, the genetic diversity of dispersing larvae must be compared with the postdispersal stages, that is benthic recruits and adults. Such data remain, however, scarce due to the difficulty to sample and analyse larvae of minute size. Here, we carried out such an investigation using the marine gastropod Crepidula fornicata. Field sampling of three to four larval pools was conducted over the reproductive season and repeated over 3 years. The genetic composition of larval pools, obtained with 16 microsatellite loci, was compared with that of recruits and adults sampled from the same site and years. In contrast to samples of juveniles and adults, large genetic temporal variations between larval pools produced at different times of the same reproductive season were observed. In addition, full‐ and half‐sibs were detected in early larvae and postdispersal juveniles, pointing to correlated dispersal paths between several pairs of individuals. Inbred larvae were also identified. Such collective larval dispersal was unexpected given the long larval duration of the study species. Our results suggest that each larval pool is produced by a small effective number of reproducers but that, over a reproductive season, the whole larval pool is produced by large numbers of reproducers across space and time.     

A hierarchical Bayesian model to estimate the unobservable predation rate on sawfly cocoons by small mammals

Predation by small mammals has been reported as an important mortality factor for the cocoons of sawfly species. However, it is difficult to provide an accurate estimate of newly spun cocoons and subsequent predation rates by small mammals for several reasons. First, all larvae do not spin cocoons at the same time. Second, cocoons are exposed to small mammal predation immediately after being spun. Third, the cocoons of the current generation are indistinguishable from those of the previous generation. We developed a hierarchical Bayesian model to estimate these values from annual one‐time soil sampling datasets. To apply this model to an actual data set, field surveys were conducted in eight stands of larch plantations in central Hokkaido (Japan) from 2009 to 2012. Ten 0.04‐m² soil samples were annually collected from each site in mid‐October. The abundance of unopened cocoons (I), cocoons emptied by small‐mammal predation (M), and empty cocoons caused by something other than small‐mammal predation (H) were determined. The abundance of newly spun cocoons, the predation rate by small mammals before and after cocoon sampling, and the annual rate of empty cocoons that remained were estimated. A posterior predictive check yielded Bayesian P‐values of 0.54, 0.48, and 0.07 for I, M, and H, respectively. Estimated predation rates showed a significant positive correlation with the number of trap captures of small mammals. Estimates of the number of newly spun cocoons had a significant positive correlation with defoliation intensity. These results indicate that our model showed an acceptable fit, with reasonable estimates. Our model is expected to be widely applicable to all hymenopteran and lepidopteran insects that spin cocoons in soil.     

Determining the optimal developmental route of Strongyloides ratti: an evolutionarily stable strategy approach

Understanding the processes that drive parasite evolution is crucial to the development of management programs that sustain long-term, effective control of infectious disease in the face of parasite adaptation. Here we present a novel evolutionarily stable strategy (ESS) model of the developmental decisions of a nematode parasite, Strongyloides ratti. The genus Strongyloides exhibits an unusual developmental plasticity such that progeny from an individual may either develop via a direct (homogonic) route, where the developing larvae are infective to new hosts, or an indirect (heterogonic) route, where the larvae develop into free-living, dioecious adults that undergo at least one bout of sexual reproduction outside the host, before producing offspring that develop into infective larvae. The model correctly predicts a number of observed features of the parasite's behavior and shows that this plasticity may be adaptive such that pure homogonic development, pure heterogonic development, or a mixed strategy may be optimal depending on the prevailing environmental conditions, both within and outside the host. Importantly, our results depend only on the benefits of an extra round of reproduction in the environment external to the host and not on benefits to sexual reproduction through the purging of deleterious mutation or the generation of novel, favorable genotypes. The ESS framework presented here provides a powerful, general approach to predict how macroparasites, the agents of many of the world's most important infectious diseases, will evolve in response to the various selection pressures imposed by different control regimes in the future.     

The role of medium conditioning in the population dynamics of the housefly

The role of larval medium conditioning on the behavior of larvae and of ovipositing adults of the housefly, Musca domestica L., was investigated through behavioral preference tests. Larvae were strongly attracted to medium conditioned by themselves or other larvae, while adult females overwhelmingly oviposited in fresh medium. Medium conditioning occurs within a few hours after hatching of an egg cohort and is effective in preventing overcrowding of a single site by shutting off further oviposition within 24 hours after eggs are initially deposited. A model of medium conditioning optimizes density for developing larvae and could also provide for regulation of local larval populations.     

Intraspecific variation in beech scale populations and in susceptibility of their host Fagus sylvatica

ABSTRACT.

• 1 Within a seed orchard in southern England, beech trees (Fagus sylvatica L.) belonging to three clones were artificially infested by introducing beech scale larvae (Cryptococcus fagisuga Lind. Homoptera: Coccidae) into small cages attached to the bark.
• 2 Some larvae developed to fecund adults on trees of two susceptible clones but all failed to develop on a third, resistant clone.
• 3 Within susceptible clones, survival of larvae on individual trees was positively related to their degree of natural infestation.
• 4 Larvae deriving from several separate trees differed significantly in their ability to survive when inoculated onto trees of susceptible clones.
• 5 Five forest trees which acted as both donors of larvae and as hosts for artificial inoculation were each inoculated with larvae from all five trees.
• 6 There was significant variation in survival of inoculated larvae both between the host trees and between sources of larvae on each host.
• 7 Survival of larvae reinoculated onto their original host was significantly higher than that of larvae originating from other trees.
• 8 Fecundity of adults on the forest trees was positively correlated with the probability of inoculated larvae surviving to the adult stage.

     

Life‐history trade‐offs under different larval diets in Drosophila suzukii (Diptera: Drosophilidae)

Most larval drosophilids eat the microorganisms that develop in rotting fruit, a relatively protein‐rich resource. By contrast, the spotted‐wing Drosophila suzukii Matsumara (Diptera: Drosophilidae) uniquely develops in ripening fruit, a protein‐poor, carbohydrate‐rich resource. This shift in larval nutritional niche has led to D. suzukii being a significant agricultural pest in the U.S.A. and Europe. Although occupying a new niche may benefit a species by reducing competition, adaptation in host use may generate trade‐offs affecting fitness. To test the hypothesis that fitness trade‐offs will change with adaptation to novel larval diets, D. suzukii larval development on either a diet of a fresh, ripe blueberry (a natural host) or standard artificial Drosophila media (protein‐rich) is compared and the effect of diet on development time from egg to adult, adult body size and male wing spot area, and female fecundity is assessed. Larval development time differs, with larvae on the blueberry emerging as adults earlier than those on the artificial medium, although other fitness measures do not vary between the two diets. In addition, the faster development time on a blueberry does not trade off with body size as expected, although early fecundity is delayed in females that develop on blueberries. Thus, adaptation to a novel larval diet environment does not come at a cost to the ability to develop in protein‐rich resources.     

Stay or move: how Bt‐susceptible Helicoverpa armigera neonates behave on Bt cotton plants

Helicoverpa armigera (Hübner) (Lepidoptera: Noctuidae) larvae occasionally have been reported to survive at management threshold levels in fields of Bollgard II^® cotton, Gossypium hirsutum L. (Malvaceae). The pattern and degree of larval survival is not easily predicted but depends on the ability of first instars to establish on host plants. Experiments were conducted with Bacillus thuringiensis Berliner (Bt)‐susceptible and Bt‐resistant larvae of H. armigera to understand how physiologically Bt‐susceptible H. armigera survive on Bt cotton plants, and examine how their first meal influences survival rates. In assays using cotton plant parts, both strains of larvae displayed similar tendencies to drop‐off specific plant parts of Bt and non‐Bt cotton. However, significantly more Bt‐susceptible larvae dropped off young leaves, mature leaves, and squares of Bt cotton compared to non‐Bt cotton plants. Egg cannibalism significantly improved the survival of Bt‐susceptible H. armigera larvae on Bt cotton plants. Larvae were more likely to eat live aged eggs, than newly laid or dead eggs. Survival significantly improved when larvae cannibalized eggs before feeding on Bt leaves. The behavior of Bt‐susceptible larvae with respect to drop‐off and egg cannibalism may help enhance their survival on Bt cotton plants.     

Toward a practical use of Neotropical odonates as bioindicators: Testing congruence across taxonomic resolution and life stages

Odonates are suggested as bioindicators of human impact. However, their complex life cycles add additional challenges in the practical use as bioindicators, because the level of taxonomic identification could be dependent on life-history stage and, during their ontogeny, dramatic changes occur in their niche (ontogenetic niche shifts). Considering that larvae and adults have different biological characteristics, which could interfere in their performance as bioindicators, we first sought to understand how similar or different environmental factors affect larval and adult life stages in the Odonata. Second, we assessed the level of congruence between (larvae and adult) and within (adult genera and species) life-history stages, considering the taxonomic and numerical resolution. We sampled larvae and adults in 44 streams distributed along a riverine network in southwest Brazil. Larvae samples constituted 20 sampling units of 1 m length each, using the kick sampling method; adults were collected for 1 h at each site with a hand net along a 100-m transect parallel to the stream/river banks. The influence of environmental factors on larvae and adult was assessed by redundancy analysis coupled with forward selection. The congruence level between response matrices was determined by Procrustes analysis. Our results revealed that a set of environmental variables explained a portion of larvae and adults distribution and some environmental factors affect both between (larvae and adults) and within (adult genera and species) life-history stages. Larvae and adult were about 54% congruent, regardless of taxonomic level of adults. Abundance of adult genera and species were 94% congruent, but numerical resolution (abundance vs. incidence) decreased the congruency by 10%. Environmental variables could influence larvae and adults individually or via carry-over effects, i.e., larval environmental conditions that could affect adult fitness components or vice versa. In addition, some odonate behaviors, such as female selection of more appropriate habitats for laying their eggs, could also help us to explain our results, because it could determine larvae distribution. In a biomonitoring perspective, considering the cost-benefit of taxonomic level and sampling of larvae and adults, our results suggest that abundance of adult genera could be used in biomonitoring programs since they capture, respectively, 94% and 54% of the information carried by adult species and larvae.     

Effect of mating status and age of Microplitis rufiventris (Hym., Braconidae) females on the growth pattern and number of their teratocytes

The effect of mating status and age of the female of Microplitis rufiventris parasitoid on the growth pattern of its teratocytes during the larval development is described as well as the changes in the number of these cells. The growth pattern of teratocytes derived from haploid eggs significantly differed from that of eggs deposited by mated females. Both cell diameter and the number of maturing M. rufiventris teratocytes was fertilization-dependent. Following the parasitization of Spodoptera littoralis larvae with virgin parasitoid females, the host larvae hosted a large number of cells of smaller size at the end of parasitoid development. The opposite effect was seen when the host larvae contained female parasitoid larvae. The age of the female at parasitization had an effect on the maximum cell size attained prior to parasitoid emergence. It seems that there are two factors other than mating which affect the number of teratocytes: degeneration inside the host haemolymph and to some degree accidental ingestion by parasitoid larva.     

Larval cannibalism during the late developmental stages of a facultatively gregarious encyrtid endoparasitoid

• 1 The larvae of many gregarious parasitoid species are usually non‐aggressive when they develop in or on a host, but those of Metaphycus flavus are one of the few exceptions known. Herein we describe their aggressive behaviour and the conditions under which it occurs, using observations in which larval development and physical conflict within parasitised and superparasitised hosts were mapped daily.
• 2 Metaphycus flavus larvae often engaged in physical conflict that resulted in consumption of the losing larvae (= cannibalism ) in superparasitised hosts, whereas such conflict and consumption occurred rarely when a single brood developed in a host.
• 3 Cannibalism among M. flavus larvae only occurred after the host resources had become scarce. Typically it occurred after the sixth day of development (fourth‐instar larvae) when the larvae in a clutch had separated from their aeroscopic plate and were freed of their attachment to the host's cuticle.
• 4 Female larvae in the initial clutch appeared more aggressive than male larvae when a second clutch was allocated 4 h after the first clutch. The probability of a larva being attacked and consumed by a brood mate increased as the number of larvae increased in the host. This partial tolerance might allow the members of the initial brood to defend themselves from offspring of a superparasitising female (= competitors ). Such post‐ovipositional regulation of brood size might be interpreted as high‐density intolerance among female offspring.