A test of temporal and spatial density dependence in the parasitism rates of introduced parasitoids on host, the arrowhead scale (Unaspis yanonensis) in stable host–parasitoids system

Abstract: In 1980, two species of parasitoid wasps (Aphytis yanonensis DeBach et Rosen and Coccobius fulvus Compere et Annecke) were introduced to Japan from China as biological control agents to combat the arrowhead scale (Unaspis yanonensis Kuwana). These introductions represent one of the most successful projects in the history of biological control in Japan. To examine whether density dependent parasitism was inevitable for success of biological control, we tried to detect temporal and spatial density dependence in parasitism rates using time‐series data of scale density, as well as parasitism, over a 16‐year period. The work was conducted in a Satsuma mandarin orange (Citrus unshiu Marc.) orchard in which we previously demonstrated that the system appeared to have stabilized after a decline in scale density following the introduction of the parasitoids. Earlier work also indicated that C. fulvus contributes most to the reduction in, and the stability of, scale density. In this study, we examined: (1) the relationship, on a whole‐orchard basis, between scale density and the rates of parasitism by A. yanonensis, C. fulvus, and a combination of the two species; (2) whether parasitism was positively correlated to scale density on a single‐tree basis among generations and (3) whether spatial density dependence was detectable within generations on an individual‐tree basis. Parasitism by A. yanonensis was temporally density‐dependent on scale population density at the whole‐orchard level, while parasitism by C. fulvus was not. Parasitism by A. yanonensis or by C. fulvus was rarely positively correlated to scale density at the single‐tree level, and spatial density‐dependence was hardly detected at all at this level. Most analyses of combined parasitism rates were similar to rates of parasitism by C. fulvus alone. Contrary to conventional wisdom of biological control theory, this study demonstrates that density dependence is not necessarily detected, even in a system in which a natural enemy has long held pest density stable at low levels.     

Introduction of parasitoids has maintained a stable population of arrowhead scales at extremely low levels

We previously reported the drastic decline of the arrowhead scale, Unaspis yanonensis Kuwana (Hemiptera: Diaspididae) following the introduction of two parasitoid species, Aphytis yanonensis DeBach et Rosen and Coccobius fulvus Compere et Annecke (Hymenoptera: Aphelinidae), which were used as biological control agents in a Japanese grove of Satsuma mandarin oranges, Citrus unshiu Marc. (Rutaceae). In this study, we examined whether the parasitoids regulated the scale population at lower levels after its initial decline. Specifically, we monitored the population dynamics of the scale and the rates of parasitism by the two parasitoids three times per year for 16 years following the introduction of the wasps. The two parasitoid species maintained a U. yanonensis density at 1/200 of the density prior to their introduction. When we excluded the wasps, the scale population grew at a rate that was more than fivefold that of a control (parasitoid‐infested) group. Although the rates of parasitism by C. fulvus fluctuated, they remained at relatively high levels, whereas those of A. yanonensis were 0% over the last 6 years. A repeated‐measures ANOVA indicated that scale density remained stable subsequent to its rapid decline. This showed that the parasitoids stabilized the scale population at a lower level than host plant limitations would have dictated, and strongly suggests that C. fulvus alone regulates the scale population density at an extremely low level. The latter finding contradicts other studies which have suggested that the two parasitoid species complement each other in regulating scale density. We discuss whether a behavioral refuge used by the scale against parasitoids, which we have demonstrated in an earlier study, might contribute to the observed stable host–parasitoid system at low densities.     

Is one parasitoid enough? A test comparing one with a pair of parasitoid species in the biological control of arrowhead scales

In classic biological control using natural enemies, the question of whether a single species or multiple species should be introduced has been a matter of debate. The introduction of two parasitoids, Aphytis yanonensis and Coccobius fulvus (Hymenoptera: Aphelinidae), to control the arrowhead scale, Unaspis yanonensis (Hemoptera: Diaspididae), which is a serious pest in Japanese citrus orchards, has been one of the most successful biological control projects in Japan. The success of this program may be explained by two alternative hypotheses: (1) the parasitoid species work complementarily, or (2) only one of them plays a major role. To test which hypothesis is applicable to this host-parasitoid system, we conducted caging experiments and observed temporal changes in the proportion of the parasitisms and the densities of arrowhead scales enclosed with one of the following combinations of parasitoids: (1) A. yanonensis and C. fulvus together, (2) A. yanonensis alone, (3) C. fulvus alone, or (4) neither parasitoid. Parasitisms in the cohorts with A. yanonensis and C. fulvus together and C. fulvus alone rapidly increased to approximately 70%; parasitism with A. yanonensis alone also increased slightly, although it remained consistently lower that those with A. yanonensis and C. fulvus together and C. fulvus. At the end of the experiment, parasitisms with A. yanonensis and C. fulvus together and C. fulvus alone were significantly higher than that with A. yanonensis alone. Parasitism by C. fulvus constituted most of (74%) the parasitism in the cohort with A. yanonensis and C. fulvus together. Further, only C. fulvus suppressed the population growth rates of scales significantly. These results suggest that C. fulvus alone successfully suppresses scale populations as efficiently as both species together do.     

Fitness cost of parasitoid avoidance behavior in the arrowhead scale, Unaspis yanonensis Kuwana

We examined trade‐offs between fitness components and the parasitoid avoidance behavior of the arrowhead scale, Unaspis yanonensis Kuwana (Homoptera: Diaspididae). We examined the performance of two settling modes: singleton and double settling with the burrower under the burrowee. Each of these modes differs in vulnerability to parasitism by Aphytis yanonensis DeBach et Rosen and Coccobius fulvus Compere et Annecke (Hymenoptera: Aphelinidae), and performance. Field and laboratory observations revealed smaller body size and lower parasitism of burrower scales, while both singletons and burrowees showed the opposite tendency (larger body size and a greater parasitism rate). Laboratory observations under parasitoid‐free conditions revealed greater mortality and reduced growth rate for simulated burrowers than for singletons during the 40‐day nymphal period. This suggests that competition for food resources reduces performance. The results also suggest trade‐offs between parasitoid avoidance and resource acquisition ability in U. yanonensis. The trade‐offs also explain our previous findings, that introduction of the parasitoids increased the proportion of burrowers, but that within seven years equilibrium was reached at fairly low levels (ca. 5%) in the field, even under intense parasitism.     

Fitness cost of parasitoid avoidance behavior in the arrowhead scale, Unaspis yanonensis Kuwana

We examined trade-offs between fitness components and the parasitoid avoidance behavior of the arrowhead scale, Unaspis yanonensis Kuwana (Homoptera: Diaspididae). We examined the performance of two settling modes: singleton and double settling with the burrower under the burrowee. Each of these modes differs in vulnerability to parasitism by Aphytis yanonensis DeBach et Rosen and Coccobius fulvus Compere et Annecke (Hymenoptera: Aphelinidae), and performance. Field and laboratory observations revealed smaller body size and lower parasitism of burrower scales, while both singletons and burrowees showed the opposite tendency (larger body size and a greater parasitism rate). Laboratory observations under parasitoid-free conditions revealed greater mortality and reduced growth rate for simulated burrowers than for singletons during the 40-day nymphal period. This suggests that competition for food resources reduces performance. The results also suggest trade-offs between parasitoid avoidance and resource acquisition ability in U. yanonensis. The trade-offs also explain our previous findings, that introduction of the parasitoids increased the proportion of burrowers, but that within seven years equilibrium was reached at fairly low levels (ca. 5%) in the field, even under intense parasitism.     

Rapid change in the settling behavior of the arrowhead scale Unaspis yanonensis as an avoidance mechanism against introduced parasitoids, Aphytis yanonensis and Coccobius fulvus

Abstract Parasitoids are thought to exert immense selection pressures that shape the traits of herbivores. We examined whether two species of parasitoid wasps, Aphytis yanonensis DeBach et Rosen and Coccobius fulvus Compere et Annecke (Hymenoptera: Aphelinidae), affect the settling behavior of the arrowhead scale Unaspis yanonensis Kuwana (Hemiptera: Diaspididae), in order to demonstrate the evolution of antiparasitism behavior of herbivores using parasitoids in the field. We used the following five methods: a comparison of parasitism risk in different settling behaviors‐parasitoid introduction into a parasitoid‐free population; a comparison of the settling behavior between parasitoid‐present and parasitoid‐free populations; a common garden experiment, in which scales were transferred from parasitoid‐present and parasitoid‐free populations into the same garden; and a laboratory observation of the settling behavior of the first instars derived from the two population categories. Both parasitoids were introduced into a parasitoid‐free population in Wakayama in 1987, and the settling mode of the scales was examined in 1987, 1994, and 1995. The introduction of parasitoids modified the scale‐settling mode so that more crawlers settled under another scale (called burrowing), which was consistent with the results observed in parasitoid‐present (including South China) and parasitoid‐free populations. Moreover, only the burrowing scale exhibited a lower parasitism rate compared to scales settling singly and being burrowed. The common garden experiment demonstrated that scales introduced from the parasitoid‐present population had a greater proportion of burrowers than the parasitoid‐free population, even in the same field cage under parasitoid‐free conditions. Laboratory observations demonstrated that the population difference in parasitism rate was principally due to intrinsic differences in the settling behavior of nymphs; some first‐instar nymphs derived from the parasitoid‐present population burrowed under another scale settled. These results strongly suggest that the natural selection pressure imposed by the parasitoids modified the settling behavior of the arrowhead scale.     

Spatio‐temporal variation in Helicoverpa egg parasitism by Trichogramma in a tropical Bt‐transgenic cotton landscape

• 1 Understanding the spatio‐temporal dynamics of insects in agroecosystems is crucial when developing effective management strategies that emphasize the biological control of pests.
• 2 Wild populations of Trichogramma Westwood egg parasitoids are utilized for the biological suppression of the potentially resistant pest species Helicoverpa armigera (Hübner) in Bt‐transgenic cotton Gossypium hirsutum L. crops in the Ord River Irrigation Area (ORIA), Western Australia, Australia.
• 3 Extensive, spatially‐stratified sampling during a season of relatively high Trichogramma abundance found that spatial patterns of pest egg parasitism in the ORIA tend toward heterogeneity, and do not necessarily coincide with host spatio‐temporal dynamics. Both patterns of host egg density and mean rates of parasitism are not good indicators of parasitoid spatio‐temporal dynamics in ORIA cotton crops.
• 4 Parasitism rates can be significantly higher within the middle strata of the cotton plant canopy before complete canopy closure, despite a similar number of host eggs being available elsewhere in the plant.
• 5 Spatial variation in egg parasitism by Trichogramma in Bt‐transgenic cotton is evident at the between‐field, within‐field and within‐plant scale, and is not solely driven by host spatial dynamics. These factors should be considered when estimating Trichogramma impact on pest species during biological control and spatio‐temporal studies of host‐parasitoid interactions in general.

     

Perspective of practical biological control and population theories

We have not yet had sufficient theoretical explanation for successful biological control in which a key pest is controlled after an introduction of natural enemies. I compare here real features of successful biological control and theoretical host–parasitoid population models to reduce the gap between theory and practice. I first review the historical interaction between classical biological control projects and theoretical population models. Second, I consider the importance of host refuges in host–parasitoid population dynamics as concerns the mechanisms of low and stable host density. The importance of density–dependent parasitism through parasitoid reproduction in multivoltine host–parasitoid systems and supplemental generalist natural enemies are also discussed. Finally, I consider the difference in tactics for classical biological control and for augmentation of natural enemies in annual crop systems. Received: December 20, 1998 / Accepted: January 15, 1999     

A hierarchical multi‐scale analysis of the spatial relationship between parasitism and host density in urban habitats

Studies on spatial density dependence in parasitism have paid scarce attention to how changes in host density at different hierarchical scales could influence parasitism in an herbivore at a particular scale. Here, we evaluated if rates of parasitism per leaf (by the whole parasitic complex and by dominant species) of the specialist leaf miner Liriomyza commelinae (Diptera: Agromyzidae) respond to variations in host density at the leaf, plant patch and site levels in an urban setting. We used multi‐level Bayesian models that incorporate the spatial hierarchy occurring in this system, as well as habitat factors previously found to have an effect on the L. commelinae parasitoid community in an urban context (patch size, patch isolation and urbanization level). According to the fitted model, overall parasitism rates decreased with increasing number of mines per leaf, being independent of host‐density variations at patch and site level. Patch structure was found to have a strong effect on parasitism rates per leaf. The analysis of parasitism by parasitoid species separately showed consistent results with the response at community level. These results suggest that parasitism of the parasitoid community here studied would be sensitive to hierarchical cues related to the host at the leaf level and to the host habitat at the patch level.     

Life cycle and biological control of the Eucalyptus snout beetle (Coleoptera, Curculionidae) by Anaphes nitens (Hymenoptera, Mymaridae) in north-west Spain

1 Gonipterus scutellatus is a weevil that has become a pest in most Eucalyptus plantations in Africa, America and Europe. The egg parasitoid Anaphes nitens has been introduced into many countries as a biological control agent. Even if control has been successful in most countries no detailed study of the interactions between both species has been published. 2 Gonipterus scutellatus was detected in 1991 in north-west Spain and A. nitens was introduced in early 1994. Here we report on the results of a 2-year study of parasitism in a field plot and a survey of 16 localities in North-west Spain. In 1996, parasitism was so intense (80–100% of eggs) that G. scutellatus became locally extinct, and as a consequence A. nitens disappeared. In 1997, G. scutellatus recolonized the area and was followed by its parasitoid, but parasitism was low in spring, probably because the parasitoid population needed 3 weeks to achieve a similar size as in 1996. Consequently, damage to the trees was extreme in 1997. We interpret these results as population fluctuations due to parasitoid–host interactions and suggest that parasitoids should not to be so effective as to locally extinguish their host to be useful for biological control. 3 The analysis of parasitism level in 16 localities indicates that A. nitens is highly efficient in finding G. scutellatus egg-masses. At a small spatial scale (single trees) positive density dependence was detected where parasitism was low and inverse density dependence where parasitism was high.     

The biology of Phymastichus coffea LaSalle (Hymenoptera: Eulophidae) under field conditions

The coffee berry borer (CBB) Hypothenemus hampei (Ferrari) (Coleoptera: Curculionidae: Scolytinae) was accidentally introduced into México in 1978, and rapidly became the main pest of coffee. As an exotic pest, its management has been mainly based on biological control methods through the introduction of parasitoids from Africa. In this context, at the beginning of the present decade, the parasitoid Phymastichus coffea LaSalle (Hymenoptera: Eulophidae) was imported to Mexico. Since then, several studies have been carried out as part of the post introduction evaluation of this parasitoid. In this paper, information concerning the parasitism and life-cycle of P. coffea in coffee farms is presented with the objective of providing information that elucidates its role as a biological control agent. P. coffea showed highly significant preferences for allocation of two eggs per host, usually one female and one male. Both offspring are able to develop and reach the adult stage successfully. Lifespan of adults is 2–3 days only. The degree of parasitism by P. coffea was more than 95% at the three altitudes tested, when releases consisted of a ratio of 10 CBB:1 parasitoid. The median survivorship of CBB parasitized by this wasp was 13, 15 and 19 days at the low, medium and high altitude coffee zones, respectively. The parasitism by P. coffea was higher when parasitoid releases were carried out simultaneously with the CBB, and decreased with the time between host and parasitoid releases. We showed that using P. coffea at a density of 1 parasitoid per 10 hosts resulted in a 3- to 5.6-fold decrease in CBB damage to the coffee seeds when compared to the control. The importance and value of these results are discussed in terms of the use of P. coffea as a biological control agent of the CBB in Latin America.     

A Simulation model for the arrowhead scale (Hemiptera: Diaspididae) population dynamics on citrus trees in relation to pest-management programs

1. A mathematical model was constructed describing population dynamics of the arrowhead scale, Unaspis yanonensis Kuwana. The population occurrence patterns simulated by the model from 1971 to 1976 were relatively consistent with the actual findings in the field.
2. The effects of control tactics for this pest were examined on the basis of the model.
3. Effective timing of insecticide (petroleum oil and organophosphorous insecticide) application was indicated to be 40 days after the initial appearance of 1st-instar nymphs in both the 1st and 2nd generations of the host.
4. The number of overwintering adult females in May of the next year (i.e., the starting point of the next year's population) decreased linearly with the increment in the degree of insecticide coverage. The pest populations were expected to be kept to low densities by petroleum oil sprays alone for the 1st generation when overall, complete coverage was accomplished.
5. Examination of the effects of two introduced parasitoids, Aphytis yanonensis and Coccobius fulvus, showed that they were able to control the scale population by themselves when more than 70% parasitism was achieved. When petroleum oil was applied to overwintering adults (termed winter petroleum oil) at the degree of coverage of 0.8, host populations were reduced by ca. 60% parasitism. When the winter petroleum oil was combined either with additional petroleum oil 40 days after the initial appearance of 1st-instar nymphs of the 1st generation or with an organophosphorous insecticide 40 days after the initial appearance of 1st-instar nymphs of the 2nd generation (degree of coverage=0.8 in each case), 40% parasitism led the reduction of the pest population.
6. The most desirable control program was considered to be one in which winter petroleum oil was used every year and additional petroleum oil or an organophosphorous insecticide was also used in those years when host density was high.

     

Spatial density‐dependent parasitism and specificity in the robber fly Mallophora ruficauda (Diptera: Asilidae)

The density‐dependence in parasitism by the robber fly Mallophora ruficauda (Diptera: Asilidae) on scarab beetle larvae (Coleoptera: Scarabaeidae) populations was studied in the present research. Mallophora ruficauda is a pestiferous species common in the open grasslands of the Pampas region of South America. Adults are predators of insects and larvae are solitary parasitoids of third instar larvae of several species of scarab beetle (Coleoptera: Scarabaeidae). In contrast with most studied host‐parasitoid interactions, host searching by M. ruficauda is carried out by both larvae and adults. Typically, robber fly females lay eggs on tall grasses from where larvae drop to the ground, and attack hosts which are buried in the soil. We carried out our study at two spatial scales close to 14 apiaries located in the provinces of Buenos Aires and Entre Ríos (Argentina). We found that parasitism is density‐independent at the larger spatial scale and inversely density‐dependent at the smaller one. We also found that M. ruficauda selects Cyclocephala signaticollis among several scarab beetle species. Specificity is observed both at large and small spatial scales. We discuss the implications of both host specificity and host searching behaviour on the observed parasitism patterns.     

Warm springs reduce parasitism of the cereal leaf beetle through phenological mismatch

Variation in weather among years may affect biological control of insect pests by influencing how well matched in phenology specialist parasitoids are with their pest hosts. A 10‐year study in western North America (Utah) revealed greater change with warm versus cool springs in the life cycle timing of the cereal leaf beetle (CLB), Oulema melanopus (L.), than of its principal enemy, the parasitoid wasp Tetrastichus julis (Walker). The beetle laid eggs, and larval populations developed in crop fields earlier on a calendar‐day basis, but nonetheless after more degree‐days had accumulated, in warmer than in cooler springs. The phenology of parasitism by wasps, in contrast, varied little among springs in relation to accumulated degree‐days. Consequently, in warmer springs, larval phenology of the CLB was delayed relative to adult parasitoid activity, and parasitism was reduced. Presently, a significant degree of biological control of the CLB results from parasitism by T. julis. By promoting phenological mismatch between host and parasitoid, however, a warming climate could weaken this biological control of the insect pest.     

Density dependence and environmental factors affect population stability of an agricultural pest and its specialist parasitoid

Host–parasitoid dynamics are intrinsically unstable unless the risk of parasitism is sufficiently heterogeneous among hosts. Spatial aggregation of parasitoids can contribute to this heterogeneity, stabilising host–parasitoid population dynamics and thereby reducing pest outbreaks. We examined the spatial distribution of mango gall fly (Procontarinia matteiana, Kiefer and Cecconi), a non-native pest of South African mango orchards, which is controlled by a single parasitoid (Chrysonotomyia pulcherrima, Kerrich). We assessed whether spatial aggregation of parasitoids is associated with proximity to natural vegetation and/or to host density-dependent and host density-independent factors at three spatial scales. We found evidence for higher parasitism rates near natural vegetation at the field scale, and inverse host-density dependent and density-independent parasitoid aggregation at both the leaf scale and field scale. Therefore, we conclude that natural vegetation plays a role in promoting stabilising aggregation of parasitoids, possibly through provision of non-host resources (nectar, pollen), in this system.     

The relationship between variable host grouping and functional responses among parasitoids of Antispila nysaefoliella (Lepidoptera: Heliozelidae)

Our study investigated the importance of variability in the parasitoid community as a source of selection on host group size using a field population of the tupelo leafminer, Antispila nysaefoliella Clemens, which specializes on tupelo, Nyssa sylvatica Marsh. Larvae were collected from leaves with variable numbers of larvae and screened for parasitism using polymerase chain reaction of mitochondrial cytochrome oxidase I using markers designed specifically for amplifying parasitoid DNA while excluding host DNA. This method of selective PCR was effective for detecting the presence and identifying species of immature stages of three hymenopteran superfamilies: Chalcidoidea, Ichneumonoidea and Platygastroidea, which represented 83.4%, 16.0% and 0.6% of the total detectable parasitism, respectively. Our resulting sequences were then calibrated with sequences from identified adult parasitoids that had been either reared or field‐captured. A cluster analysis revealed 10 distinct clades that showed differences in attack patterns with respect to host traits and season. Total parasitism followed an inverse density‐dependent or density‐independent pattern with respect to host density (number per leaf). However, when parasitoid taxa were considered separately, one clade, which could be a cryptic species of Pnigalio maculipes Crawford (Chalcidoidea: Eulophidae), was found to increase its per leaf attack rate with host density. Our results suggest that parasitoid community composition and differences among species in their attack strategies can play a large role in determining the adaptive advantage of host grouping.     

Do distances among host patches and host density affect the distribution of a specialist parasitoid?

The effect of spatial habitat structure and patchiness may differ among species within a multi-trophic system. Theoretical models predict that species at higher trophic levels are more negatively affected by fragmentation than are their hosts or preys. The absence or presence of the higher trophic level, in turn, can affect the population dynamics of lower levels and even the stability of the trophic system as a whole. The present study examines different effects of spatial habitat structure with two field experiments, using as model system the parasitoid Cotesia popularis which is a specialist larval parasitoid of the herbivore Tyria jacobaeae. One experiment examines the colonisation rate of the parasitoid and the percentage parasitism at distances occurring on a natural scale; the other experiment examines the dispersal rate and the percentage parasitism in relation to the density of the herbivore and its host plant. C. popularis was able to reach artificial host populations at distances up to the largest distance created (at least 80 m from the nearest source population). Also, the percentage parasitism did not differ among the distances. The density experiment showed that the total number of herbivores parasitised was higher in patches with a high density of hosts, regardless of the density of the host plant. The percentage parasitism, however, was not related to the density of the host. The density of the host plant did have a (marginally) significant effect on the percentage parasitism, probably indicating that the parasitoid uses the host plant of the herbivore as a cue to find the herbivore itself. In conclusion, the parasitoid was not affected by the spatial habitat structure on spatial scales that are typical of local patches.     

Spatial and temporal correspondence of sunn pest oviposition and egg parasitism by Trissolcus spp. in Iran

The sunn pest, Eurygaster integriceps Puton, is a serious pest of wheat and other cereals in the Firouzabad region of Ardabil Province, Iran. In this study, we examined the spatial distribution of sunn pest eggs and their parasitism by Trissolcus spp., primarily Trissolcus grandis Thomson, during the growing seasons of 2008 and 2009. Ten quadrats of 1 m² were established in each of the ten experimental fields and sampled every 3 days beginning at Feekes growth stage 6 (beginning of stem elongation) through stage 11 (ripening). The numbers of sunn pest eggs, both parasitized and unparasitized, were tallied in each quadrat on each sampling date. Sunn pest oviposition peaked during the onset of flowering in both years, while parasitism peaked at the completion of anthesis in 2008, and at anthesis complete to early milk stage in 2009. Iwao’s regression provided a better fit to the data than did Taylor’s power law. Both models indicated an aggregated distribution of eggs and parasitism. Temporally, there was a negative overlap between egg density and parasitism from inflorescence emergence until the beginning of anthesis, and positive overlap from anthesis complete until the soft dough stage, such that parasitism ranged from 0% of eggs laid early, to virtually 100% of eggs laid late. Thus, despite good spatial correspondence between egg density and egg parasitism, the time lag between peak bug oviposition and peak parasitism enabled large numbers of early‐laid host eggs to hatch and diminished the potential efficacy of biological control by Trissolcus spp. The implications of these results are discussed in the context of integrated management of sunn pest in wheat fields and the possible benefits of early season augmentation of the parasitoid population.     

Stability in the host-parasitoid relationship of Dialeurodes citri and Encarsia lahorensis in the citrus orchard

The citrus whitefly, Dialeurodes citri (Ashmead), a citrus pest, has been prevalent in Israel since 1975. The parasitic wasp Encarsia lahorensis (Howard) has been successfully used for its biological control since 1980 and thehost-parasitoid system is stable. This studyuses field data from four and a half years inorder to examine whether heterogeneity ofparasitism and risk aggregation can explain thestability. After establishing that theprobability of parasitism is not constant overpatches, we examined the question of parasitoidaggregation, dependent or independent of thehost, at different patch levels. At the treelevel we found an inverse relationship betweenthe proportion of parasitism and host density.At the leaf level, taking the tree effect intoconsideration, the host density dependence wasweak and non-significant. At the leaf level, acombined examination of both types ofheterogeneity in parasite distribution – hostdensity dependent heterogeneity (direct orindirect) and host density independentheterogeneity, was performed using the (CV)² > 1 criterion. The mean (CV)² value over different sampling dateswas greater than one. Host density independentheterogeneity had the greater contribution tostability. The (CV)² findings atleaf level in the plot, combining leaf and treeeffects, represent both aggregation at the treelevel (inverse density dependence) andaggregation at the leaf level (host densityindependence). The heterogeneity findings forparasitoid distribution, together with thestability, which was directly observed, supportour hypothesis that aggregation of risk is animportant mechanism in the stabilisinginteraction between the citrus whitefly and itsspecific parasitoid, E. lahorensis.     

Spatial patterns of Trybliographa rapae parasitism of Delia radicum larvae in oilseed rape and cauliflower

Trybliographa rapae (Westwood) is an important parasitoid of Delia radicum (L.). Parasitism of D. radicum larvae by T. rapae in relation to host density on canola (oilseed rape) and cauliflower roots was examined at 10 field sites in Germany and Switzerland. For roots with host larvae, the proportion of roots with one or more parasitized hosts increased with increasing host density. However, for these infested roots, the parasitism of individual larvae was not consistently related to host density. When considering only roots on which there were parasitized larvae and the opportunity for multiple attacks, the proportion of larvae that were parasitized decreased with increasing host density in the field locations, and in a cage study under controlled conditions. A model of patch‐finding and number of attacks by female parasitoids suggests that patch‐finding is density‐dependent, but that low attack rate and interference effects limit numbers of attacks to three or less per visit to a host patch; the reduced number of attacks per visit leads to the inverse relationship of larval parasitism with host density in the host patches visited. The interplay of the density‐dependent and inversely density‐dependent processes appears to be responsible for the inconsistency of density dependence of overall larval parasitism in this and previous studies. In the laboratory, adult female T. rapae parasitized hosts at ≤4 cm deep in soil, but not at 6 cm deep. From the depth distribution of larval feeding sites in the field, we infer that between 4% and 20% of Delia larvae may be in a physical refuge from T. rapae parasitism, which may have a stabilizing influence on the host–parasitoid interaction.