Optimal selection and exploitation of hosts in the parasitic wasp Colpoclypeus florus (Hym., Eulophidae)

This study deals with the question how an insect parasitoid can maximize its fitness through adaptation of its reproductive behaviour. It concentrates on the behaviour of a parasitoid after it has encountered a host. Optimal exploitation of individual hosts is emphasized rather than a maximization of the number of parasitized hosts.

In Chapter I the topic of optimization of behaviour is introduced in relation to the study of insect parasitoids. The choice of the experimental animals is explained and behavioural alternatives of the parasitoid are discussed. In this study the number of granddaughters is taken as a measure of fitness.

The chalcidoid wasp Colpoclypeus florus (Hym., Eulophidae) is a gregarious ectoparasitoid of larvae of at least 32 species of leafrollers (Lep., Tortricidae; Table 2.1). Host plants are predominantly trees and shrubs. The parasitoid has a west palearctic distribution (Fig. 2.1) and is rare in natural or semi-natural habitats. However, C. florus can be found in abundance in intensively cultivated habitats. In the Netherlands they are found especially in apple orchards, during outbreaks of the summer fruit tortrix moth, Adoxophyes orana . Efforts to control A. orana with mass releases of the parasitoid had not been successful. However, the parasitoid is considered as promising by those working on integrated control and more biological information was required.

In Chapter 2 the parasitization behaviour, development and phenology of the parasitoid is described. The experimental host ( A. orana ), general techniques and conditions are also described. Field experiments were carried out in an experimental apple orchard. Unlike many internal and external parasitoids, C. florus has the unusual habit of ovipositing beside instead of on or in the host. This offers the opportunity to manipulate the eggs and hosts separately. In addition, the number of hosts parasitized by an individual in the field is low, about 2-3 hosts per female, and the time taken to parasitize one host is long (average 13-28 h in the laboratory at 21 °C and about twice as long in the field, in summer). Thus, C. florus is particularly suitable for studies on how it optimizes exploitation of individual hosts.

Three stages in the parasitization process were analysed in detail.

(a) The first problem concerned the host size selection for oviposition (Chapter 3 and 4). It was hypothesized that only the most profitable hosts are selected for oviposition. Only the first of five larval instars of A. orana is rejected for oviposition by the parasitoid. In the laboratory, proportion of hosts accepted, clutch size, survival of pre-adults, proportion females and parasitization time increase with host weight (Tables 3.3 and 3.5). As a result the profitability of hosts (defined as the fitness gained per unit of time or per egg) is correlated with host acceptance, but the profitability threshold of host acceptance is low (Fig. 3.2). It was shown that this threshold is not influenced by changes in the length of the pre-encounter period with hosts. C. florus is assumed to be unable to measure host density, therefore size preference may be genetically fixed and be an adaptation to low host densities. Host acceptance is the same in the laboratory and the field. In the field, however, nearly all parasitized hosts are fourth and fifth larval instars (Table 5.5). It was shown that the parasitoids during the host-habitat location phase have a preference for the young leaves in the outer layer of the canopy where the larger hosts predominate (Fig. 4.4). These larger hosts are also more conspicuous to the parasitoids during host location. The combined effect of host-habitat location and host location could explain why in the field few small hosts are parasitized resulting in a host size selection favouring the most profitable host sizes (Table 4.7). Assuming a genetically fixed optimal host choice, a first theoretical estimate of the host encounter rate in the field was made.

(b) The second question dealth with how many eggs were to be laid near each selected host (Chapter 5). It was hypothesized that the clutch size would maximize the profitability from a selected host. The parasitoid was found to adapt her clutch size to the weight of the host, apparently using a factor related to the host width. This results in a curvi-linear relationship between host weight and clutch size (Fig. 5.2). Experimentally the number of eggs per host was manipulated, and the number and individual weights of offspring were determined. Although pre-adult mortality in the laboratory, as well as in the field is high (50-60 %), density dependent mortality of juveniles does not occur in the normal clutch sizes (Fig. 5.16). However, competition for food among the juveniles always occurs, which results in body weights attained not being maximal (Fig. 5.17). Longevity and fecundity of a female are positively related to her weight (Fig. 5.12 and 5.14). Thus the clutch size per host size ratio affects total longevity and fecundity of the offspring. Two strategies are discussed. The first requires a low number of eggs per host and results in maximum offspring fecundity per egg laid. The second requires a high number of eggs per host and results in maximum offspring longevity per parasitized host. It was shown that C. florus produces a clutch size per host size ratio which cannot be explained by either of these strategies. It is assumed that female parasitoids will obtain just that body weight which will allow her to invest all her eggs during her lifetime. With this assumption, a second theoretical estimate of the host encounter rate in the field was made. This second estimate falls within the range of that assuming optimal host choice. Therefore, host size selection and clutch size per host size ratio are both optimal in the same range of host encounter rate.

(c) The third problem concerned that of sex allocation (Chapter 6). The sex ratio of adult C. florus is female biased and the proportion of males decreases with increasing clutch size (Table 6.5). The number of males per clutch is constant, or may slightly increase with increasing clutch size. During development males and females do not suffer differential mortality (Table 6.3). The insemination capacity of males increases with increasing body weight (Fig. 6.2). A male has sufficient time available and, assuming that he has obtained a mean body weight, he has just enough insemination capacity to inseminate all sisters of his clutch even in the largest clutches. It is suggested that fitness of males cannot be increased by a higher body weight, since females outside the clutch are seldom encountered. Using a newly developed cytological technique, it was demonstrated that male eggs are laid at the end of an ovipositional sequence and not at random throughout egg laying (Fig. 6.4). This points to a precise sex ratio mechanism. A model assuming such a mechanism, and using the information that one adult male from a clutch of eggs is sufficient to inseminate all females, accurately predicted the actual number of unfertilized eggs in a clutch (Fig. 6.6). It is suggested that in species where host size is clearly limited and is estimated by the parasitoid before oviposition males are allocated at the end of a clutch.

In Chapter 7 the main conclusions are given and discussed. This study gives insight into how a parasitic wasp tackles the different difficulties which arise in optimizing its reproductive behaviour. Although constraints operate at different levels of the parasitization process, clear examples of optimal behaviour (as defined in this study) are found in C. florus . For an understanding of some aspects of the behaviour, the information collected about the field situation appeared to be essential.