Journal of Insect Behavior, VoL 7, No. 4, 1994
Host-Searching Behavior of Venturia c a n e s c e n s (Grav.) (Hymenoptera: Ichneumonidae): Superparasitism J. P. Hughes, 1'2 I. F. Harvey, 1'3 and S. F. Hubbard 1'4
Accepted July 20, 1993; revisedJanuary 18, 1994
The foraging behavior of Venturia canescens, a solitary endoparasitoid of lepidopteran larvae, was investigated in the laboratory. Female Venturia canescens with a larger number of mature eggs to lay were found to have higher levels of superparasitism (measured as numbers of eggs laid per parasitized host). Increased parasitoid density was found to result in reduced levels of superparasitism by host-deprived (i.e., undepleted) wasps. Females which had been allowed access to hosts before the experiment (depleted wasps) laid fewer eggs per parasitized host than undepleted wasps, although there was no significant difference in the levels of superparasitism among the depletion periods of 1, 2, 5, and 7 h. It was also found that an egg which was encountered less than 15 min after oviposition was much less likely to be avoided than one which was encountered after more than 15 rain had elapsed. KEY WORDS: superparasitism; egg load; host-parasitoid system; foraging; Venturiacanescens.
INTRODUCTION F o r many years, superparasitism by solitary parasitoids was regarded as evidence o f occasional deficiencies in their sensory capabilities, resulting in the wastage ~Department of Biological Sciences, University of Dundee, Dundee DDI 4HN, UK. 2present address: Division of Environmental and Evolutionary Biology, Bute Building, University of St. Andrews, St. Andrews KY16 9TS, UK. 3present address: Population Biology Research Group, Department of Environmental and Evolutionary Biology, University of Liverpool, P.O. Box 147, Liverpool L69 3BX, UK. 4"I"o whom correspondence should be addressed at Department of Biological Sciences, The University, Dundee DD1 4HN, UK.
455 0892-755319410700-0455507.0010 © 1994 Plenum Publishing CoqJoratlon
456
Hughes, Harvey, and Hubbard
of both eggs and time. Under most conditions it does seem likely that it will be advantageous for a searching parasitoid to refrain from ovipositing in a previously parasitized host. More recently, however, a number of theoretical (Chamov and Skinner, 1984, 1985; Iwasa et al., 1984; Parker and Courtney, 1984; Skinner, 1985; Mangel, 1992) and experimental (Bakker et al., 1985; Hubbard et al., 1987; Marris, 1988; Visser et al., 1990) studies have raised the possibility that it may, in certain circumstances, pay a parasitoid to remain in a host patch and oviposit into already parasitized hosts. These workers have approached superparasitism as an optimal diet or optimal patch use problem, where a female behaves in a way that maximizes her fitness gain per unit time, and not simply her oviposition rate. The models predict that superparasitism should increase at lower rates of encounter with healthy hosts (i.e., high rates of encounter with previously parasitized hosts) or if the probability of the second or subsequent egg laid winning the ensuing contest is greater than zero. A second approach has been to use the methods of game theory to analyze superparasitism, since superparasitism involves a response to the behavior of other females (van Alphen and Visser, 1990; Hoeven and Hemerik, 1990; Visser et al., 1990, 1992). Iwasa et al. (1984) pointed out that, if handling time and risk of mortality are the same for both parasitized and unparasitized hosts, the decision whether or not to superparasitize will be strongly influenced by the number of mature eggs available for oviposition, with parasitized hosts being avoided more frequently by wasps with a smaller egg load. For a solitary parasitoid, superparasitism of one of her own eggs (selfsuperparasitism) would, in most circumstances, not be adaptive since the competition within the host would be between full sibs, of which only one would survive. Superparasitism in a host containing an egg laid by a conspecific female, however, could prove to be an adaptive strategy if, as stated above, the probability of the offspring from the second egg winning and surviving to adulthood is greater than zero. This is so, particularly if the transit times between patches of hosts are long relative to the total available searching time or involve a high risk of mortality (van Alphen, 1988). Visser and co-workers (1992) have developed an ESS model of patch depletion which predicts that no superparasitism will be found in patches depleted by a single female and that, in patches depleted by more than one female, the degree of superparasitism will increase with the number of females. Qualitative agreement with these predictions was found in an experiment using the eucoilid, Leptopilina heterotoma (Visser et al., 1990). Hubbard et al. (1987) and Marris (1988), working with the ichneumonid Venturia canescens, have shown that, as one would expect if the wasp distinguishes between self- and conspecific superparasitism, females discriminate
Superparasitism, Egg Load, and Interference in Parasitoids
457
between hosts containing their own progeny and those containing the progeny of conspecifics. They do this by means of a relatively short-lived chemical marker, originating from Dufour's gland, which is released into the host at oviposition. In addition, Hubbard et al. (1987) also demonstrated that female V. canescens will superparasitize significantly more often in those hosts parasitized by a conspecific. Similarly, Bakker et al. (1985) and Visser et al. (1990) found that Leptopilina heterotoma females, eucoilid parasitoids of Drosophilid larvae, searching in the presence of conspecific females readily superparasitize, whereas those searching a host patch alone do not. From such results it can be seen that an adaptive patch use strategy is not necessarily simply to reduce the rate of encounter with healthy hosts on the patch to the average rate of encounter with healthy hosts for the environment as a whole and then depart, as proposed by Cook and Hubbard (1977). If, for instance, transit times between patches are relatively long and the offspring from the second or subsequent egg(s) laid has a greater than zero chance of winning the intraspecific contest within the host, then the adaptive strategy might well be to remain on the patch for longer if other females are present and lay additional eggs. Such a strategy could result in mutual interference between searching females (Visser et al., 1992). For a fuller review of adaptive superparasitism, see the excellent review by van Alphen and Visser (1990). In this paper, we report on the effect of egg load and parasitoid density on the pattern of oviposition and superparasitism exhibited by the parasitoid, Venturia canescens (Grav.) (Hymenoptera: Ichneumonidae). This species is especially useful in studies of this type because it performs a characteristic movement of the ovipositor and abdomen after the deposition of an egg, the so-called "cocking" movement (Rogers, 1972), allowing the observer to distinguish between an oviposition and an unsuccessful host encounter without dissecting hosts after experiments. MATERIALS AND METHODS A full description of the culturing procedures for the parasitoid, Venturia canescens, and its phycitid host, Plodia interpuncteUa (Hubner) (Lepidoptera: Pyralidae), and the experimental procedure is given in the preceding paper (Hughes et al. 1994). RESULTS For each wasp in each triM, the mean number of eggs per parasitized host (EPH) was calculated, and subsequent analysis is based on such means. First we examine, by three-way analyses of variance, the effects of the depletion period, parasitoid density, and depletion opportunity on EPH (Table
458
Hughes, Harvey, and Hubbard
I). Overall, both density and depletion opportunity have significant effects, but the length of the depletion period does not. Additionally, none of the interactions are significant. However, these relationships are not simple. In Fig. 1 it is shown that the undepleted wasps show no significant change in EPH at densities of 2, 4, or 8 wasps, but that EPH falls dramatically when the parasitoid density is increased to 16 wasps. To investigate the effects of starting egg load, we conducted a one-way analysis of covariance on EPH with parasitoid density as the main effect and egg load as a covariate. The null hypothesis of homogeneity of slopes for the relationship between egg load and EPH within each parasitoid density could be accepted (F = 0.60, df = 3,251, P = 0.605), and the pooled regression coefficient was significantly different from zero (F = 24.41, df = 1,254, P < 0.001). So egg load has a significant effect on EPH, the relationship being slight (pooled r 2 = 9.4%) and positive.
Table I. The Effect of Parasitoid Density, Length of Depletion Period, and Depletion Oportunity on the Number of Eggs Laid per Parasitized Host by V. canescens Females (Three-Way ANOVA) Source
df
Main effects Depletion time (A) Parasitoid density (B) Depletion opportunity (C) Interaction A × B A x C B x C A x B x C Error Total
La'
SS
MS
F
P
3 3 1
0.2564 1.0759 3.2062
0.0855 0.3686 3.2062
0.62 2.68 23.34
ns <0.05 <0.001
9 3 3 9 230 261
2.2453 0.2553 0.8978 1.2211 12.8539 15.5094
0.2495 0.0851 0.2993 0.1357 0.1037
1.82 0.62 2.18 0.99
ns ns ns ns
t"10e#eted
I"1Undepleted
1.6' EPH
/.4' 12' l 1.02
8 Parasitoiddensity
16
Fig. 1. The effect of parasitoid density on numbers of eggs laid per parasitized host (EPH) by Venturia canescens. Values are means :I:SE. Data pooled for depletion periods. Numbers in bars are sample sizes.
Superparasitism, Egg Load, and Interference in Parasitoids
459
Parasitoid density, controlling for among parasitoid variations in egg load, also affects EPH (ANCOVA: F = 3.16, df = 3, 254, P = 0.025). If the observed differences between depleted and undepleted wasps are the result of differences in starting egg load alone, then a one-way ANCOVA on differences in EPH between wasp types, controlling for egg load, should reveal no significant differences in EPH. Pooling data for parasitoid density, depletion time, and wasp type, we find that the pooled regression is significantly different from 0 (F = 8.55, df = 1,256, P = 0.004), and there are significant differences between wasp types (ANCOVA: F = 7.95, df = 1,256, P = 0.005). The mean EPHs, adjusted to take account of differences in starting egg load, are as follows: depleted wasps, 1.298 (SD = 0.3465); and undepleted wasps, 1.441 (SD = 0.3152). So the length of the depletion period has no significant effect on the observed levels of superparasitism, simply having been depleted has an effect over and above that of egg load reduction. From the data gathered for the two-parasitoid trials, it was possible to establish directly the rates of avoidance for both wasp types, when encountering their own and their competitor's eggs. The proportion of multiple encounters (encounters with hosts which have already been parasitized at least once) in which total, self- and conspecific superparasitism was avoided was calculated for each wasp in each of the two-parasitoid trials. For this analysis, all instances where the host encountered already contained both self-laid and conspecific eggs were excluded from the analysis, since it is impossible to say which egg is being avoided. Rogers (1972) demonstrated that the avoidance rate of V. canescens reaches a maximum between 15 and 20 min after oviposition. So the results were also examined for any difference in the various avoidance rates both less and more than 15 min after the first oviposition. The mean rates of avoidance of superparasitism are given in Table II. We conducted a two-way ANOVA to investigate the effects of wasp type (depleted or undepteted), egg type (self or conspecific), and egg age ( < 15 or > 15 min) on the rate of avoidance of superparasitism. No interaction terms were significant, and neither was the effect of egg type (F = 0.003, df = 1,147, P = 0.954). However, both egg age and wasp type were significant (wasp type, F = 8.98, df = 1,147, P = 0.003; egg age, F = 23.75, df = 1,147, P < 0.001). Depleted wasps avoided superparasitism more frequently than undepleted wasps (Mann-Whitney U test, P < 0.05) (see Fig. 2), and for both wasp types avoidance rates were significantly higher when the first-laid egg was encountered more than 15 min after it was laid. To investigate the effects of egg load we calculated correlations between avoidance of superparasitism and egg load. Pooling data for self- and conspecific superparasitism, it was found that, although there are trends in these data show-
460
Hughes, Harvey, and Hubbard Table H. The Effects of Depletion Opportunity and Egg Age
on Superparasitism: Mean Rates of Avoidance of Superparasifism (+ 1 SE) Superpamsitism Self < 15 min >15 rain Conspecific < 15 min >15 min
Depleted
Undepleted
49.8 ± 8.25 (n = 21) 76.6 5:4.63 (n --- 19)
33.2 + 5.04 (n = 32) 62.5 + 5.29 (n = 32)
59.4 + 7.66 (n = 18) 73.4 5:6.56 (n = 23)
33.3 + 9.59 (n = 19) 62.3 -I- 7.73 (n = I5)
ing that avoidance rates were higher at high egg loads, there was no significant correlation for depleted (r 2 = 0.022) or undepleted (r 2 = 0.013) wasps. This result reinforces the earlier finding that the effect o f the depletion period on the results is not due simply to the reduction in egg load. This result was confirmed by a three-way A N C O V A , with egg load as the covariate. Egg load had no significant effect on the avoidance o f superparasitism, and inclusion o f the covariate made very little change in the results o f the three-way A N O V A outlined above. DISCUSSION As pointed out by Rosenheim and Rosen (1991), the effects o f egg load and experience are likely to be difficult to differentiate. Our attempts to manipulate egg load by depletion have resulted in the creation o f two groups o f parasitoids with different perceptions o f host availability. The results clearly show that having previous experience o f oviposition into healthy hosts has a direct effect on the behavior o f V. c a n e s c e n s females distinct from the effect o f egg load. That egg load is influential in the decision to superparasitize is shown in our results. Other studies employing similar "host-deprivation" techniques have failed to resolve the relative influences o f experience and egg load (e.g., Ikawa and Suzuki, 1982; Fitt, 1986, 1990; Pilson and Rausher, 1988; Putters and van den Assem, 1988; Volkl and Mackauer, 1990). This study shows that in V. c a n e s c e n s both factors have a role, but more experimental work is required to clarify the relative importance o f these factors. The results obtained also suggest that the undepleted wasps are unwilling to superparasitize at the highest parasitoid/host ratio. What reasons could there
Superparasitism, Egg Load, and Interference in Parasitoids
461
80"
Avoidance Rate (%)
40'
20'
0 DEPLETED
i/
UNOEPLETED
WaspState Fig. 2. Comparison of rates of avoidance of superparasitism for depleted and undepleted wasps at parasitoid density of two wasps. Values are means +SE. Sample sizes: depleted wasps, N = 30; undepleted wasps, N = 35.
be for this finding? Although at lower densities there may be some reward gained by engaging in a "war of attrition" (Maynard Smith, 1982; van Alphen, 1988) with conspecifics because of a greater-than-zero probability of the second or subsequent eggs laid winning the ensuing larval competition, it is reasonable to assume that there will be some upper threshold beyond which the reward drops to zero. This could occur, for example, if the hosts suffer an increase in mortality through injuries received during oviposition, or, if a greater number of puncture wounds rendered a host more susceptible to bacterial or viral infection. Another explanation is that such superparasitism as does occur at high parasitoid-host ratios may occur because the available hosts are soon parasitized and at least some of these are likely to be reencountered before the chemical marker (Hubbard et al., 1987; Marris, 1988) has become apparent (i.e., during the period less than 15 min after the initial oviposition). There is no significant change in the number of eggs per parasitized host with increasing depletion time, although the lowest value is found after 7 h. This lends support to previous findings, that simply being depleted is the important factor, and not the actual length of the depletion period, at least on the time scale of these experiments. Depletion has been shown to have a similar effect on searching and fighting behavior (Hughes, 1988; Hughes et al., 1994). This experiment is similar to that proposed by Marc Mangel (1989) to test the hypothesis that wasps deprived of hosts will have a higher acceptance rate of previously parasitized hosts than wasps of the same age allowed access to hosts. There is, obviously, good support for Mangel's hypothesis in our results. In this study, there were instances of repeated oviposition into certain hosts, a few of which received as many as eight eggs in total. Explaining findings such as these is difficult. However, there are a few possible explanations, which
462
Hughes, Harvey, and Hubbard
are dependent on the possibility that second and subsequent eggs can win the larval competition, van Lenteren (1976) has shown that the eucoilid Leptopilina heterotoma has the ability to distinguish between hosts with difering numbers of eggs. If Fenturia is capable of the same, then it may be that females are attempting to ensure that the eventual victor is one of their progeny, responding to superparasitism by opponents with the laying of further eggs. Instances of what appear to be tit-for-tat sequences of oviposition into particular hosts were observed in these experiments. Bakker et al. (1985) provide some experimental evidence, using wild-type and mutant strains of L. heterotoma, of the larvae from second eggs laid surviving the competition. They also found that success in such contests was dependent on the interval between the successive ovipositions. In addition to these competition explanations, there is the possibility that the first egg laid into a host may be encapsulated by the host immune response, leaving the host unable to encapsulate any eggs laid subsequently. Observed encapsulation rates in this host species range from 0 to 100%, and the phenomenon is, as yet, only poorly documented and understood. All of these possible explanations need to be examined experimentally, and the fitness payoffs from single and multiple parasitism quantified, if we are to be able to explain and predict patterns of superparasitism in Venturia canescens. Perhaps even more importantly we require field data from the range of vastly different habitats in which V. canescens is found if we are to unravel the selective pressures which shape the wasp's behavior. In this paper we deal with a situation in which wasps are searching the patch simultaneously. However, as Visser et al. (1992) have shown, superparasitism behavior may well be different if wasps visit patches sequentially. ACKNOWLEDGMENTS Thanks to Wallace Arthur, Jeremy Greenwood, and Giles Thomas for statistical advice and to Marcel Visser for useful comments on the manuscript. We would also like to thank the following for practical help of various kinds: Willie Black, Georgina Green, Alice Ruthven Hughes, John Johnston, Gay Marris, Neil McCuUoch, and Jonathan Stewart. This work was supported by the NERC, the SERC and the Royal Society. REFERENCES Bakker, K., van Alphen, J. J. M., van Batenburg, F. H. D., van der Hoeven, N., Nell, H. W., van Strien-van Liempt, W. T. F. H., and Turlings, T. C. J. (1985). The function of host discrimination and superparasitization. Oecologia 67: 572-576. Charnov, E. L., and Skinner, S. W. (1984). Evolutionof host selection and clutch size in parasitoid wasps. Fla. Entomol. 67: 5-21.
Superparasitism, Egg Load, and Interference in Parasitoids
463
Charnov, E. L., and Skinner, S. W. (1985). Complementary approaches to the understanding of parasitoid oviposition decisions. Environ. Entomol. 14: 383-391. Cook, R. M., and Hubbard, S. F. (1977). Adaptive searching strategies in insect parasites. J. Anim. Ecol. 46: 115-125. Fitt, G. P. (1986). The influence of a shortage of hosts on the specificity of oviposition behaviour in species of Dacus (Diptera: Tephritidae). Physiol. Entomol. 11: 133-143. Fitt, G. P. (1990). Comparative fecundity, clutch size, ovariole number and egg size of Dacus tryoni and D. jarvisi, and their relationship to body size. Entomol. Exp. Appl. 55: 11-21. Hubbard, S. F., Marris, G. C., Reynolds, A., and Rowe, G. W. (1987). Adaptive patterns in the avoidance of superparasitism by solitary parasitic wasps. J. Anim. Ecol. 56: 387--401. Hughes, J. P., Harvey, I. F., and Hubbard, S. F. (1994). Host-searching behavior of Venturia canescens (Gray.) (Hymenoptera: Ichneumonidae): Interference--The effect of mature egg load and prior behavior. J. Insect Behav. 7: 433--454. Ikawa, T., and Suzuki, Y. (1982). Ovipositional experience of the gregarious parasitoid, Apanteles glomeratus (Hymenoptera: Ichneumonidae), influencing her discrimination of the host larvae, Pieris rapae crucivora. Appl. Entomol. Zool. 17:119-126. Iwasa, Y., Suzuki, Y., and Matsuda, H. (1984). Theory of oviposition strategy of parasitoids. I. Effect of mortality and limited egg number. Theor. Popul. Biol. 26: 205-227. Mangel, M. (1989). An evolutionary interpretation of the "motivation to oviposit." J. Evol. Biol. 2: 157-172. Mangel, M. (1992). Descriptions of superparasitism by optimal foraging theory, evolutionary stable strategies and quantitative genetics. Evol. Ecol. 6: 152-169. Marris, G. C. (1988). Studies on the avoidance of Superparasitism in the Parasitic Hymenoptera, Unpublished Ph.D. thesis, University of Dundee, Dundee, UK. Man-is, G., Hubbard, S., and Hughes, J. (1986). Use of patchy resources by Nemeritis canescens (Hymenoptera: Ichneumonidae). L Optimal solutions. J. Anim. Ecol. 55: 631-640. Maynard Smith, J. (1982), Evolution and the Theory of Games, Cambridge University Press, Cambridge, UK. Parker, G. A., and Courtney, S. P. (1984). Models of clutch size in insect oviposition. Theor. Popul. Biol. 26: 27-48. Pilson, D., and Rausher, M. D. (1988). Clutch size adjustment by a swallowtail butterfly. Nature (Lond.) 333: 361-363. Putters, F. A., and van den Assem, J. (1988). The analysis of partial preferences in a parasitic wasp. Anim. Behav. 36: 933-948. Rogers, D. J. (1970). Aspects of Host-Parasite Interaction in Laboratory Populations of Insects, Unpublished D.Phil. thesis, University of Oxford, Oxford, UK. Rogers, D. J. (1972). The ichneumon wasp Venturia canescens: Oviposition and avoidance of superparasitism. Entomol. Exp. Appl. 15: 190-194. Rosenheim, J. A., and Rosen, D. (1991). Foraging and oviposition decisions in the parasitoid Aphytis lignanensis: Distinguishing the influences of egg load and experience. J. Anim. Ecol. 60: 873-893. Skinner, S. W. (1985). Clutch size as an optimal foraging problem for insects. Behav. Ecol. Sociobiol. 17: 231-238. van Alphen, J. J. M. (1988). Patch time allocation by insect parasitoids: Superpamsitism and aggregation. In de Jong, G. (ed.), Population Genetics and Evolution, Springer-Verlag, Berlin/ Heidelberg, pp. 215-221. van Alphen, J. J. M., and Visser, M. E. (1990). Superparasitism as an adaptive strategy. Annu. Rev. Entomol. 35: 59-79. van der Hoeven, N., and Hemerik, L. (1990). Superparasitism as an ESS: To reject or not to reject, that is the question. J. Theor. BioL 146: 467-482. van Lenteren, J. C. (1976). The development of host discrimination and the prevention of superparasitism in the parasite Pseudeucoila bochei Weld (Hym.: Cynipidae). Neth. J. Zool. 26: 1-13. Visser, M. E., van Alphen, J. J. M., and Nell, H. W. (1990). Adaptive superparasitism and patch
464
Hughes, Harvey, and Hubbard
time allocation in solitary parasitoids: The influence of the number of parasitoids depleting a patch. Behaviour 114: 21-36. Visser, M. E., van Alphcn, J. J. M., and Hemerik, L. (1992). Adaptive superparasitism and patch time allocation in solitary parasitoids: An ESS model. J. Anita. Ecol. 61: 93-101. Volkl, W., and Mackancr, M. (1990). Age-specific pattern of host discrimination by the aphid parasitoid Ephedrus californicus Baker (Hymenoptera: Aphidiidae). Can. Emomol. 122: 349361.