Chapter 2. Life history variation in aphid hyperparasitoids: Is development mode a major determinant?

Table des matières

2.1. Abstract

In primary parasitoids, development mode has been identified as a major determinant of life histories. The dichotomous hypothesis contrasts the life history traits of koinobiont endoparasitoids with those of idiobiont ectoparasitoids. In this study we examined if the dichotomous hypothesis can also be applied to hyperparasitoids, or if the hyperparasitic strategy demands different adaptations that will confound the dichotomous hypothesis. We compared life history parameters of two koinobiont endoparasitic species ( Alloxysta victrix (Westwood) and Syrphophagus aphidivorus (Mayr)) and two idiobiont ectoparasitic species ( Asaphes suspensus (Nees) and Dendrocerus carpenteri (Curtis)) of aphid hyperparasitoids from four different families that attack either the parasitoid larva in the aphid before it is killed and mummified by the primary parasitoid, or the parasitoid prepupa or pupa in the dead aphid mummy.

The variation in life history traits in aphid hyperparasitoids cannot be explained by development mode alone. The data for the idiobiont ectohyperparasitoids mostly confirm the dichotomous hypothesis. The koinobiont endohyperparasitoids, however, have a long adult lifespan and a low fecundity, contrary to the predictions of the dichotomous hypothesis. These traits are best explained by synovigeny in these species. It is likely that several factors, including development mode, timing of egg production, host range and host stage, act together and are selected to optimise fitness. In addition, lineage specific effects might also determine life history traits.

Keywords : parasitoid, dichotomous hypothesis, koinobiont, idiobiont, host range, host stage, phylogeny.

2.2. Résumé

Chez les parasitoïdes primaires, le mode de développement à été reconnu comme facteur déterminant de l’histoire de vie. L’hypothèse de la dichotomie contraste les caractéristiques de l’histoire de vie des endoparasitoïdes koinobiontes et des ectoparasitoïdes idiobiontes. Dans cette étude nous avons examiné si l’hypothèse de dichotomie peut aussi être appliquée aux hyperparasitoïdes ou si l’hyperparasitisme exige des adaptations qui la confondent. Nous avons comparé les paramètres de l’histoire de vie de deux endo-hyperparasitoïdes koinobiontes ( Alloxysta victrix (Westwood) et Syrphophagus aphidivorus (Mayr)) et deux ecto-hyperparasitoïdes idiobiontes ( Asaphes suspensus (Nees) et Dendrocerus carpenteri (Curtis)), provenant de quatre familles différentes. Ces espèces attaquent soit la larve du parasitoïde dans le puceron avant qu’il soit tué, soit la prépupe ou la pupe du parasitoïde dans la momie de puceron.

La variation observée des caractéristiques d’histoire de vie des hyperparasitoïdes de puceron n’a pu être expliquée par le mode de développement seul. Les résultats pour les ecto-hyperparasitoïdes idiobiontes étaient en accord avec l’hypothèse de dichotomie pour la plupart des paramètres. Par contre, les endo-hyperparasitoïdes koinobiontes avaient une vie adulte plus longue et une fécondité plus basse, que prédit par l’hypothèse. Ces caractéristiques sont mieux expliquées par la synovigénie chez ces espèces. Il est probable que divers facteurs, incluant le mode de développement, la précocité de l’ovogénèse, la spécificité parasitaire et le stade d’hôte, agissent ensemble et sont sélectionnés pour optimiser le fitness. En plus, des effets d’ordre phylogénétique peuvent aussi déterminer les caractéristiques d’histoire de vie.

2.3. Introduction

Life histories across a range of organisms have been studied extensively to identify what determines the values of life history traits (Stearns, 1992; Roff, 1992). Within the parasitic Hymenoptera, development mode (koinobiont or idiobiont) was proposed as the major organiser of life history patterns in parasitoids (reviewed by Godfray, 1994; Quicke, 1997; Mayhew and Blackburn, 1999; Strand, 2000). Koinobiont parasitoids allow further development of the host after parasitism, and are mostly endoparasitic. This enables them to attack exposed, early instar hosts that have less efficient defences against parasitism.Koinobiosis implies a slow or delayed larval development, small eggs and high fecundity, and a short adult lifespan. Idiobiont parasitoids attack non-growing host stages such as eggs or pupae or arrest host development via the injection of venoms or other biochemical factors preceding oviposition. It is to their advantage to attack large, mature host stages as their host represents a fixed amount of resources. Therefore, idiobiont parasitoids are expected to possess a different (and potentially opposite) set of life history characteristics when compared to koinobionts. The dichotomous hypothesis states that natural selection operates on the life history strategies of these two categories of parasitoids to magnify the differences (Godfray, 1994). The first test of the dichotomous hypothesis was performed by Mayhew and Blackburn (1999) on life history data of 474 species of parasitoid Hymenoptera. Only partial support for the dichotomy hypothesis was found, as life history traits were not all correlated to development mode. Still, they concluded that the support was sufficient to retain development mode as the cental element of any comprehensive theory of parasitoid life histories, especially for parasitoids of larval hosts. Development mode probably reflects the evolutionary history of parasitoid lineages while other parameters, like fecundity, are not as strongly correlated to life history traits (Mayhew and Blackburn, 1999). Other factors have been proposed to explain the same variation in life history of parasitoids, such as the degree of pro-ovigeny or synovigeny (Jervis et al ., 2001). These factors appear to be linked to parasitoid development mode so the effects of development mode and ovigeny are potentially difficult to separate.

Hyperparasitoids (or secondary parasitoids) parasitise the immature stages of primary parasitoids and therefore belong to the fourth trophic level in many ecosystems. Hyperparasitism has a wide taxonomic distribution, suggesting diverse evolutionary origins (Gordh, 1981). Among the Hymenoptera, hyperparasitism occurs in 7 of 11 parasitoid superfamilies. It is mostly found in the superfamilies Ceraphronoidea, Chalcidoidea, Ichneumonoidea, and Trigonalyoidea (Brodeur, 2000). Because of their common evolutionary origins, hymenopterous parasitoids and hyperparasitoids share many biological characteristics. Like parasitoids, ectoparasitism is generally associated with idiobiont development and endoparasitic hyperparasitoids are mostly koinobiont. Therefore these developmental traits might also be used to explore interspecific variation in life history traits in hyperparasitoids. In this study, we examined if the dichotomous hypothesis can also be applied to hyperparasitoids, or if being a hyperparasitoid demands specific adaptations that confound the dichotomous hypothesis. Based on the great similarities in development mode between primary and secondary parasitoids, we predicted that we will find the same dichotomy in life history parameters according to development mode in hyperparasitoids.

The most intensive studies of hyperparasitism have been conducted on the Hymenoptera that attack immature parasitoids developing in Homopteran hosts. Aphid hyperparasitoids are the best known group of hyperparasitoids in terms of taxonomy, host association, development mode, behaviour and impact on primary parasitoid populations (Sullivan, 1987, Mackauer and Völkl, 1993). However, a detailed and accurate comparison among species to determine the influence of development mode on life history variation is not yet possible. Much of the published information is incomplete or anecdotal and often originates from different aphid-primary parasitoid systems whose host and host plant species can vary in suitability. Therefore, in our experiments, we adopted a comparative approach in which we compared life history traits of four different aphid hyperparasitoids reared on the same primary parasitoid host species. This permits us to generalise, suggests hypotheses and places intra-specific patterns into context (Stearns, 1992). We chose one species from each Hymenoptera family that contains aphid hyperparasitoids, with the exception of the Eulophidae (Table 2-1): Two idiobiont ecto-hyperparasitoids, Asaphes suspensus (Pteromalidae) and Dendrocerus carpenteri (Megaspilidae), and two koinobiont endo-hyperparasitoids, Alloxysta victrix (Charipidae) and Syrphophagus aphidivorus (Encyrtidae).

Besides variations in development modes, these four species also differ in host range, from oligophages like A. victrix to generalists like A. suspensus , as described by Höller et al. (1993) (Table 2-1). Furthermore, they attack their immature host at different stages of development. Either they attack the parasitoid larva in the still living aphid, or they attack the parasitoid (pre-)pupa in the dead, mummified aphid (Table 2-1). These are ecological factors that have to be taken into account in the comparison because they also may influence life history traits, and may confound the results that are expected based on the dichotomous hypothesis.

We compared in the laboratory the most important life history parameters of the four hyperparasitoid species (survival, developmental time, size, longevity, fecundity, immature mortality and sex ratio) and the intrinsic rate of natural population increase (rm) of each species. In the discussion, we also used data from the literature on related aphid hyperparasitoids.

2.3.1. Description of species

For this study the same potato – potato aphid – Aphidius nigripes system was used as the basic food web. The potato aphid ( Macrosiphum euphorbiae Thomas) is a common pest on potatoes in North America (Shands, 1965; Radcliffe, 1982). These aphids live in colonies on leafs and stems. In nature, the potato aphid reproduces asexually during the summer, giving birth to female nymphs. At the end of summer, sexual forms are produced, which migrate to their primary host plant, roses, and aphids overwinter in the egg stage on these plants. In the laboratory the colonies are easily maintained asexually. The primary parasitoid Aphidius nigripes Ashmead (Hymenoptera: Braconidae) is the dominant parasitoid of the potato aphid in North America (Walker et al ., 1984). It is a solitary koinobiont species and the females attack the aphid in the nymphal or adult stages (Cloutier et al , 2000). The parasitoid larval instars feed within the aphid, ultimately killing their host. Before completing its development, the mature larva (also called prepupa) spins a cocoon inside the empty aphid cuticle and the typical aphid “mummy” is formed (Stary, 1988). The parasitoid pupa develops within the mummy.

All tested hyperparasitoid species are solitary, recorded from both Europe and North America, and are all found naturally attacking A. nigripes in the field (Shands, 1965; Brodeur and McNeil, 1994).

Asaphes suspensus (Nees) (Hymenoptera: Pteromalidae) is an idiobiont, ectophagous hyperparasitoid of aphidiine and aphelinid parasitoids, attacking the host (pre-)pupa after it has killed the aphid and the mummy is formed (Sullivan, 1987). It can be a tertiary parasitoid on its own species (Levene and Sullivan, 1983) or on other hyperparasitoids (Carew and Sullivan, 1993). The female A. suspensus envenoms its host, which deteriorates to a blackened mass on which the hyperparasitoid larva feeds (Bocchino and Sullivan, 1981). Hosts can also be used for destructive host feeding, for which the female constructs a feeding tube to feed on the host heamolymph (Levine and Sullivan, 1983; Christiansen-Weniger, 1992).

Similar to A. suspensus , Dendrocerus carpenteri (Curtis) (Hymenoptera: Megaspididae) is a solitary, idiobiont, ectophagous hyperparasitoid of aphidiine and aphelinid wasps inside mummified aphids (Fergusson, 1980; Sullivan, 1987; Chow and Mackauer, 1999). It can also be a tertiary parasitoid its own species (Bennet and Sullivan, 1978) or other hyperparasitoids (Matejko and Sullivan, 1984). Females inject their host with juvenile hormone (Höller et al ., 1994), which arrests its development (Bocchino and Sullivan, 1981). The larva feeds externally. No host feeding is reported for this species.

Alloxysta victrix (Westwood) (Hymenoptera: Charipidae) is a koinobiont endophagous hyperparasitoid that attacks parasitoid larvae in aphids before mummification (Sullivan, 1987; Gutierrez and van den Bosch, 1970). It is only after the aphid is mummified by the primary parasitoid larva, that the hyperparasitoid egg hatches and the larva starts to feed endophagously until it kills and completely consumes the host (Sullivan, 1987). The potential host range is broad (Grasswitz and Reese, 1998), but A. victrix seems to prefer Aphidiinae (Gutierrez and van den Bosch, 1970; Andrews, 1978).

Syrphophagus aphidivorus (Mayr) (Hymenoptera: Encyrtidae) is also a koinobiont, endophagous hyperparasitoid. But compared to other aphid hyperparasitoids this species is atypical because it has a dual host-stage relationship and oviposition behaviour. The female hyperparasitoid attacks either the primary parasitoid larva in an aphid when the aphid is still alive, or the parasitoid prepupa or pupa after the mummy has been formed (Sullivan, 1987). However, choice experiments indicated that mummies are preferred (Kanuck and Sullivan, 1992; Buitenhuis et al ., submitted) and data suggest that it is also from these mummies that the parasitoid gains the highest fitness (Buitenhuis et al ., submitted). When a parasitoid larva is attacked in a live aphid, the egg of the hyperparasitoid hatches only following mummification of the aphid by the host prepupa (Kanuck and Sullivan, 1992). The larva of S. aphidovorus feeds endophagously until it kills its host. Females can feed on the heamolymph that escapes from the puncture hole made by the ovipositor after oviposition (Griswold, 1929; Kanuck and Sullivan, 1992). It attacks aphidiine and aphelinid primary parasitoids (Hoffer and Stary, 1970; Sullivan and van den Bosch, 1971; Mertins, 1985; Völkl and Barczak, 1990). It might be expected that this species shows some characteristics in life history that are intermediate between hyperparasitoids that attack live parasitised aphids or those that attack aphid mummies.

2.4. Materials and Methods

2.4.1. Insect colonies

Laboratory cultures of the potato aphid ( Macrosiphum euphorbiae (Thomas)), the primary parasitoid Aphidius nigripes Ashmead and A. victrix , A. suspensus , D. carpenteri and S. aphidivorus ) were established and maintained at 20 ± 1°C, 75 ± 10% RH under a 16L:8D photoperiod. The hyperparasitoid species originated from laboratory colonies on different aphid – parasitoid systems and were therefore held in the laboratory for more than 5 generations to adapt to the potato aphid – A. nigripes system. Asaphes suspensus came from Québec City, Canada, D. carpenteri from Burnaby, Canada, A. victrix from Newport, England and S. aphidivorus from Bayreuth, Germany.

Aphidius nigripes was reared on aphid colonies feeding on potato seedlings cv. ‘Norland’ following the techniques of Brodeur and McNeil (1994). All hyperparasitoids were maintained by weekly exposing potato plants, infested with aphid mummies (for A. suspensus , D. carpenteri and S. aphidivorus ) or parasitised aphids (for A. victrix ) to the hyperparasitoid females.

2.4.2. Hosts used in the experiments

Hosts for S. aphidivorus, D. carpenteri and A. suspensus were A. nigripes prepupae in newly (0-24 h) mummified aphids (“mummy” host), while A. victrix was given 3rd instar A. nigripes larvae in living aphids as hosts (“parasitised aphid” host). In several aphid parasitoid-hyperparasitoid systems, these two parasitoid developmental stages have been shown to be the most suitable for hyperparasitoids attacking their host either when the aphid is alive (Kanuck and Sullivan, 1992; Grasswitz and Reese 1998) or once the latter has mummified (Chow and Mackauer 1999). To obtain A. nigripes hosts of both stages, third-instar potato aphid nymphs were parasitised by 3 to 5-d-old mated A. nigripes females during 24 h. Parasitised aphids were then reared at 20 ± 1ºC, 75 ± 10% RH, and a 16L:8D photoperiod. Based on embryonic and larval developmental times of A. nigripes at 20ºC (Paré et al . 1979), 3rd instars in living aphids and prepupae in newly mummified aphids were obtained 5 and 8 days following parasitisation, respectively.

All experiments were carried out in a climate chamber at 20 ± 1°C, 75 ± 10% RH and a 16L:8D photoperiod. Asaphes suspensus , D. carpenteri and S. aphidivorus were tested at the same time. For logistic reasons, A. victrix was tested 5 months later.

2.4.3. Development time and longevity

Approximately 400 hosts (parasitised aphids for A. victrix , and mummies for A. suspensus , D. carpenteri and S. aphidivorus ) on potato plants were exposed to parasitism by twenty 1 to 7 days old, mated female hyperparasitoids for 4 hours. Afterwards, mummies were collected from the plants and put in individual gelatine capsules to complete hyperparasitoid development. The parasitised aphids were left on the plants until mummification after which mummies were also collected and put in individual gelatine capsules. Temperature in gelatine capsules was checked regularly using thermocouples (OMEGA HH23). The hosts were monitored every 8 hours for hyperparasitoid emergence and the sex of each adult was recorded. Hyperparasitoids that had not emerged 10 days after peak emergence were excluded from the analysis.

To measure hyperparasitoid longevity, newly emerged hyperparasitoids from the development time experiment (n > 30 per hyperparasitoid species and sex) were kept isolated in small, ventilated cylindrical cages, (5 cm in diameter and 10 cm in height) with a supply of 40% sugar water, replaced every 3-4 days. The hyperparasitoids did not have access to hosts and were checked daily for mortality.

2.4.4. Fecundity, immature mortality and sex ratio

In a preliminary experiment, we determined the maximum daily number of hosts that each hyperparasitoid species could parasitise. Based on these results, the number of hosts that was provided daily to a female in the experiment was set to 30 mummies for D. carpenteri, A. suspensus , 70 mummies for S. aphidivorus and 30 parasitised aphids for A. victrix . These numbers insured that hosts were available ad libitum.

The females were obtained as in the development time experiment, isolated in cages in a climate chamber and provided with 40% sugar water (see longevity experiment). The first 5 days of the test, 2 males were present for mating. Every day, until death, females were given new hosts which were available for 24 h. The mummies were glued with non-toxic white glue (Lepage Bondfast®) on a potato leaf that was held by the petiole in a glass vial containing wet cotton wool and inserted into the cage. The parasitised aphids were transferred on a similar leaf with a paintbrush. After exposure to a hyperparasitoid female, mummies were put individually in gelatine capsules. Parasitised aphids were left on the leaf until mummification and then put in capsules. All capsules were held in a climate chamber until emergence. Mummies from which nothing had emerged were dissected (40X magnification) 10 days after hyperparasitoid peak emergence from a daily cohort to determine if they contained a primary parasitoid or a hyperparasitoid. Dead hyperparasitoids were classified in three categories: larva/prepupa, pupa or unemerged adult.

Realised fecundity (number of offspring that reach the adult stage), potential fecundity (all offspring, including those that died before adult emergence), pre-oviposition period (time from emergence to first oviposition), oviposition period (period during which females laid eggs) and post-oviposition period (time after last oviposition until death) were calculated. The secondary sex ratio (proportion of males) was determined at adult emergence. Mortality was expressed as the proportion of hyperparasitised hosts that contained a dead hyperparasitoid larva/prepupa, pupa or unemerged adult. Females that escaped or died by accident were excluded from the analysis. Fifteen females were tested per species.

2.4.5. Adult body size

Cohorts of parasitised aphid hosts and mummy hosts were produced and parasitised as described above (see Hosts used in the experiments). To avoid hyperparasitoid adult size being affected by mummy size, mummies of similar weight (0.9-1.1 mg) were selected. Less than 24 h after emergence, hyperparasitoids were killed by freezing at -20ºC, dried for 4 days at 60ºC and individually weighed on a Mettler Toledo UMT microbalance. In addition, measurements of the head width and front wing length (from humerus to apex) were made to the nearest 0.01 mm using a stereomicroscope (40X magnification) equipped with an ocular micrometer. At least 20 individuals were measured per species per sex.

2.4.6. Intrinsic rate of increase (rm)

The rm was estimated for each hyperparasitoid species by repeated iteration of the Birch formula (Birch, 1948):

Σ e-rmx lxmx = 1

where x is female age, lx is the fraction of females surviving to age x and mx is the age-specific fertility that records the number of living females born per female per day, calculated from the daily sex ratio data measured in the fecundity experiment.

2.4.7. Statistical analysis

Differences in development time between hyperparasitoid species and between males and females were tested with a two-way ANOVA following the Poisson distribution (GENMOD). The differences in longevity and size between between species and between males and females were tested with a standard twoway ANOVA. Immature mortality percentages and sex ratios were arcsin-transformed prior to analysis with a standard ANOVA. Potential and realised fecundity data were square-root transformed and analysed with an ANCOVA. All species are synovigenic and produce eggs throughout their life (Sullivan, 1987). This means that total fecundity is correlated with the duration of the period that females lay eggs. Oviposition period was therefore used as a covariable, Oviposition period and pre- and postoviposition periods were square-root transformed prior to analysis with a standard ANOVA. All means were separated by Fisher’s protected LSD with Bonferroni adjustment of the significance level (α=0.05/k; k=number of comparisons). All data were analysed using SAS (SAS, 1999).

2.5. Results

Values of the life history parameters and intrinsic rates of increase of the four hyperparasitoid species are summarised in Table 2-2.

2.6. Discussion

According to the dichotomous hypothesis (Godfray, 1994; Quicke, 1997; Mayhew and Blackburn, 1999), variation in life history traits can be explained by a dichotomy in development mode. However, our results based on direct comparison of four species (Table 2-2) suggest that the variation in life history traits in aphid hyperparasitoids cannot be explained by development mode alone. Some of their traits support the hypothesis, eg. koinobiosis in the tested hyperparasitoids is associated with endoparasitism, no or temporary paralysis and slow or delayed development, while idiobiosis is associated with a generally different set of traits. However, no such grouping between koinobionts and idiobionts was possible based on the life history parameters that we measured in this study. The species A. suspensus was clearly different from the other three species in longevity. Furthermore A. suspensus and S. aphidivorus both had a high fecundity and a high proportion of male offspring as opposed to A. victrix and D. carpenteri , which had lower fecundities, and more female offspring. For none of the life history parameters measured could the koinobiont species A. victrix and S. aphidivorus be separated as a group from the idiobiont species D. carpenteri and A. suspensus . It is clear that other factors, besides development mode, influence the life history traits of aphid hyperparasitoids.

Among the obtained results three things should be explained. First, in A. victrix , the mortality during the last (unemerged adult) stage of development was higher than in the other species (Table 2-2). In addition, the intrinsic rate of increase calculated based on all life history parameters of this species was the lowest of all species. Because few data are available for A. victrix , we do not know if these values are normal for this species or if the Aphidius nigripes – Macrosiphum euphorbiae system might be less suitable for the development of this hyperparasitoid species. Furthermore, this species was not tested at the same time as the other three species, which might have influenced the results. Second, the observed pattern of sex ratio is most likely associated with sperm availability. During the second half of their life, females were observed to lay only male offspring. The amount of sperm acquired during the five days that males were present might not have been enough for females to produce an optimal sex ratio as females might have run out of viable sperm long before the end of their oviposition period. It is possible that in these species females mate several times during their lifetime to replenish their sperm supply like Brodeur and McNeil (1994) proposed for the aphid hyperparasitoid Asaphes vulgaris . Finally, when we compare our results with the parameters reported in the literature, it appears that our results are equivalent or higher than those of other studies (Spencer, 1926 (cited in Schooler, 1996); Gutierrez and van den Bosch, 1970; Walker and Cameron, 1981; Christiansen-Weniger, 1992; Völkl and Kranz, 1995; Chow and Mackauer, 1996; Grasswitz and Reese, 1998). The differences are possibly due to the size of the host in our rearing system, compared to the hosts used in the other studies (various Aphidius species on Acyrthosiphon pisum (Harris), Sitobion avenae (F.), Myzus persicae (Sulzer) or Uroleucon jaceae L.). It shows that for a comparison between species it is important to rear the hyperparasitoids on the same parasitoid-aphid-plant system.

In table 2-3 the life history parameters of the four hyperparasitoids are compared to those of (primary) parasitoids in general in the context of the dichotomous hypothesis (Quicke, 1997; Mayhew and Blackburn, 1999). The expected koinobiont life history characteristics are listed on the left and those of idiobiont parasitoids to the right, as predicted by the dichotomous hypothesis. In the middle, the hyperparasitoids (on genus level to have access to more data) are compared to this model. The data were measured in this study or found in the literature. For continuous variables we compared the data on a scale between the extreme values that are known for parasitoids.

The data for the two idiobiont ectohyperparasitoids are mainly in agreement with the hypothesis. The most important exceptions are the long development time and high fecundity of A. suspensus . These traits are also found in another Asaphes species (e.g. A. vulgaris , Brodeur and McNeil, 1994). In contrast to D. carpenteri , Asaphes species can host-feed, which provides the essential nutrients to produce large eggs. For D. carpenteri , the nature of the yolk bodies and the origin of the substances used to form them are not known. The low fecundity of this species is perhaps related to the difficulty in obtaining resources required to produce yolk-rich eggs with external nutrients limited to carbohydrates, honeydew or pollen (Le Ralec, 1995). The long development time of Asaphes species might be correlated to their long lifespan. Further research has to point out if these two traits are correlated and their function in the biology of these species.

Compared to the idiobiont ectohyperparasitoids, the data for the two koinobiont endohyperparasitoids diverge much more from the predictions of the hypothesis and the results of Mayhew and Blackburn (1999). The most striking differences are the long adult lifespan and the high occurrence of egg production (low ovigeny index) in these species (Table 2-3). Furthermore, A. victrix has a very low oviposition rate (eggs/day) (mean daily fecundity 2.4 offspring, maximum 4.0 offspring per day), which is contrary to the predictions, although other species within the Alloxysta genus have somewhat higher oviposition rates (Chua, 1979; Singh and Srivastava, 1987; Mackauer and Völkl, 1993). The other endohyperparasitoid, S. aphidivorus , has the highest oviposition rate of the four studied species (mean daily fecundity 20.0 offspring, maximum 41.8 offspring per day), but these oviposition rates are still low compared to other koinobiont (primary) parasitoids (40-140 offspring per day; Aphidiidae (Force and Messenger, 1964)). Further inconsistencies with the dichotomous hypothesis for S. aphidivorus are that it is a generalist, has large eggs and is capable of host feeding. This species is able to hyperparasitise mummies and still living parasitised aphids. It strongly prefers mummies, and has higher fitness on this host (Kanuck and Sullivan, 1992; Buitenhuis et al ., submitted). So, although it is a koinobiont endohyperparasitoid, it shares more life history characteristics with idiobiont ectohyperparasitoids.

In general, it is likely that many endoparasitic koinobionts should have higher ovigeny indices than ectoparasitic idiobionts. However, in our study we observed a high level of egg production in the koinobiont hyperparasitoids, and therefore should assign them a low ovigeny index (Jervis et al ., 2001). This might explain the divergence of the results from the predictions of the dichotomous hypothesis. It appears that the low ovigeny value in these species has more influence on life history than development mode as this is correlated to long adult lifespan, large eggs and host feeding (Jervis et al ., 2001). There is some evidence that natural selection adjusts egg production characteristics to approach the expected rate of host encounter (Jervis et a l., 2001). A correlation may exist between synovigeny and a greater degree of host dispersion (Quicke, 1997). Parasitised aphids and mummies are not abundant hosts, because aphidiid wasps that oviposit in aphid colonies usually lay only a few eggs per colony and show high dispersal (Dettner et al ., 1997). However, the actual availability of hosts for hyperparasitoids has still to be elucidated.

Other factors, besides development mode, that could potentially influence life history characteristics in the hyperparasitoids of this study are host stage and host range. Because the two host stages differ in many aspects (for example morphology, olfactory cues, abundance, number of competitive species) hyperparasitism of living parasitised aphids vs. aphid mummies does not necessarily demand the same adaptations in life history. If either host stage or host range would have been the major organisers of the evolution of life history traits in aphid hyperparasitoids, we would expect A. victrix to be different from the other species, because it attacks only the parasitised aphid before mummification, and also has a narrower host range than the other three species. It is restricted to Aphidius hosts and is considered to be more specialised than A. suspensus and D. carpenteri that attack several genera within the Aphidiidae, as well as an Aphelinidae (Höller et al ., 1993). For S. aphidivorus the known host associations involve at least four primary parasitoid genera from the Aphidiidae and the Aphelinidae. However, based on the measured life history traits, A. victrix could not be placed apart from the other species. It is the closest to D. carpenteri in longevity and fecundity, while this species attacks uniquely mummified aphids, and has a large host range. In addition, we observed great differences between the species that attack mummified aphids. Both A. suspensus and D. carpenteri parasitise mummies of various parasitoids in various aphid species, and appear to occupy the same habitats and to have a large host range. However, we observed that A. suspensus lives much longer, has a higher fecundity and a longer fertile period, and has a longer development time than D. carpenteri (Table 2-2). Therefore, both host stage and host range cannot explain the differences in life history traits in these aphid hyperparasitoids.

Although the dichotomous hypothesis explains many parasitoid life history traits, it is unlikely that life history traits are determined exclusively by the dichotomy in development mode. It is clear that the variation in life history in aphid hyperparasitoids cannot be explained by single factors like development mode, synovigeny, host stage or host range. Probably all factors act at the same time on life history evolution. Darwin (1859) supposed that a balanced interpretation of an evolutionary pattern requires two components: adaptation and lineage specific effects. The effect of adaptation is that life history traits are adapted to each other and to local environmental conditions. At the same time, some life history traits are fixed at high taxonomic levels (lineage specific effects). Following this reasoning, the differences between species might also be partly determined by their different phylogenetic origins. Hyperparasitism has a wide taxonomic distribution, indicating that it has evolved independently several times in the Hymenoptera (Gordh, 1981). It is likely because the expression of hyperparasitism is phylogenetically spotty, that different development modes occur among hyperparasitoids and that hyperparasitoid wasps have different ovipositional strategies. The different species have probably evolved within the phylogenetic constraints of their origin to exploit the same resource. All the above mentioned factors act together and are selected to optimise fitness gain during the life of an individual.

2.7. Acknowledgements

We thank Dr. W. Völkl, Dr. A. Chow and Dr. T. Grasswitz, for providing hyperparasitoid species. This study was supported by a grant from the Natural Sciences and Engineering Research Council of Canada to J.B. and by graduate scholarships from the Centre de Recherche en Horticulture and Agriculture and Agri-Food Canada to R.B.

2.8. References

Andrews, F. G. 1978. Taxonomy and host specificity of Nearctic Alloxystinae with a catalogue of the world species (Hymenoptera: Cynipidae). California Department of Food and Agriculture.

Bennet, A. W., and D. J. Sullivan. 1978. Defense behavior against tertiary parasitism by the larva of Dendrocerus carpenteri, an aphid hyperparasitoid. Journal of the New York Entomological Society 86:153-160.

Birch, L. C. 1948. The intrinsic rate of increase of an insect population. Journal of Animal Ecology 17:15-26.

Bocchino, F. J., and D. J. Sullivan. 1981. Effects of venoms from two aphid hyperparasitoids, Asaphes lucens and Dendrocerus carpenteri (Hymenoptera: Pteromalidae and Megaspilidae), on larvae of Aphidius smithi (Hymenoptera: Aphidiidae). Canadian Entomologist 113:887-889.

Brodeur, J., and J. N. McNeil. 1994. Life history of the aphid hyperparasitoid Asaphes vulgaris Walker (Pteromalidae): Possible consequences on the efficacity of the primary parasitoid Aphidius nigripes Ashmead (Apidiidae). Canadian Entomologist 126:1493-1497.

Brodeur, J. 2000. Host specificity and trophic relationships of hyperparasitoids. Pages 163-183 in M. E. Hochberg and A. R. Ives, editors. Parasitoid population biology. Princeton Univ. Press.

Carew, W. P., and D. J. Sullivan. 1993. Interspecific parasitism between two aphid hyperparasitoids Dendrocerus carpenteri (Hymenoptera: Megaspilidae) and Asaphes lucens (Hyemnoptera: Pteromalidae). Annals of the Entomological Society of America 86:794-798.

Chow, A., and M. Mackauer. 1996. Sequential allocation of offspring sexes in the hyperparasitoid wasp, Dendrocerus carpenteri . Animal Behavior 51:859-870.

Chow, A., and M. Mackauer. 1999. Host handling and specificity of the hyperparasitoid wasp, Dendrocerus carpenteri (Curtis) (Hym., Megaspilidae): importance of host age and species. Journal of Applied Entomology 123:83-91.

Christiansen-Weniger, P. 1992. Wirt-Parasitoid Beziehung zwischen Blattlausprimärparasitoiden und den Blattlaushyperparasitoiden Asaphes vulgaris Wlk. und Asaphes suspensus (Nees) (Hymenoptera: Pteromalidae). PhD Thesis, Christian Albrechts University, Kiel, Germany.

Chua, T. H. 1979. A comparative study of the searching efficiencies of a parasite and a hyperparasite. Res. Popul. Ecol. 20:179-187.

Cloutier, C.; J. Duperron,; M. Tertuliano, J. N. McNeil. 2000. Host instar, body size and fitness in the koinobiotic parasitoid Aphidius nigripes. Entomologia experimentalis et applicata,. 97:29-40.

Darwin, C. 1859. The Origin of Species. John Murray, London.

Dettner, K., G. Bauer, and W. Völkl. 1997. Evolutionary patterns and driving forces in vertical foodweb interactions. in K. Dettner, G. Bauer, and W. Völkl, editors. Vertical foodweb interactions. Springer-Verlag, Berlin Heidelberg.

Fergusson, N. D. M. 1980. A revision of the British species of Dendrocerus Ratzeburg (Hymenoptera: Ceraphronoidea) with a review of their biology as aphid hyperparasites. Bulletin of the Brittish Museum of Natural History 41:255-314.

Force, D. C., and P. S. Messenger. 1964. Fecundity, reproductive rates, and innate capacity for increase of three parasites of Therioaphis maculata (Buckton). Ecology 45:706-715.

Godfray, H. C. J. 1994. Parasitoids. Behavioral and Evolutionary Ecology. Princeton University Press, Chichester, UK.

Gordh, G. 1981. The phenomenon of insect hyperparasitism and its taxonomic occurence in the Insecta. Pages 10-18 in D. Rosen, editor. The role of hyperparasitism in biological control: A symposium. Division of Agricultural Science, University of California, Berkely, California.

Grasswitz, T. R., and B. D. Reese. 1998. Biology and host selection behaviour of the aphid hyperparasitoid Alloxysta victrix in association with the primary parasitoid Aphidius colemani and the host aphid Myzus persicae . Biocontrol 43:261-271.

Griswold, G. H. 1929. On the bionomics of a primary parasite and of two hyperparasites of the geranium aphid. Annals of the Entomological Society of America 22:438-457.

Gutierrez, A. P., and R. van den Bosch. 1970. Studies on host selection and host specificity of the aphid hyperparasite Charips victrix (Hymenoptera: Cynipidae) 3. Host suitability studies. Annals of the Entomological Society of America. 63:1485-1490.

Gutierrez, A. P., and R. van den Bosch. 1970. Studies on host selection and host specificity of the aphid hyperparasite Charips victrix (Hymenoptera: Cynipidae) 6. Description of sensory structures and a synopsis of host selection and host specificity. Annals of the Entomological Society of America 63:1705-1709.

Haviland, M. D. 1920. On the bionomics and post-embryonic development of certain Cynipid hyperparasites of Aphides. Quarterly Journal of Microscopical Science 65:451-478.

Haviland, M. D. 1922. On the post-embryonic development of certain chalcids, hyperparasites of Aphides, with remarks on the bionomics of hymenopterous parasites in general. Quarterly Journal of Microscopical Science 66:321-338.

Hoffer, A., and P. Stary. 1970. A review of biologies of palearctic Aphidencyrtus - species (Hym., Chalcidoidea, Encyrtidae). Studia Entomologica Forestalia 1:81-95.

Höller, C., C. Borgemeister, H. Haardt, and W. Powell. 1993. The relationship between primary parasitoids and hyperparasitoids of cereal aphids: an analysis of field data. Journal of Animal Ecology 62:12-21.

Höller, C., H. Bargen, S. B. Vinson, and D. Witt. 1994. Evidence for the external use of juvenile hormone for host marking and regulation in a parasitic wasp, Dendrocerus carpenteri . Journal of Insect Physiology 40:317-322.

Jervis, M. A., G. E. Heimpel, P. N. Ferns, J. A. Harvey and N. A. C. Kidd. 2001. Life-history strategies in parasitoid wasps: a comparative analysis of ‘ovigeny’. Journal of Animal Ecology 70:442-458.

Kanuck, M. J., and D. J. Sullivan. 1992. Ovipositional behavior and larval development of Aphidencyrtus aphidivorus (Hymenoptera: Encyrtidae), an aphid hyperparasitoid. Journal of the New York Entomological Society 100:527-532.

Le Ralec, A. 1995. Egg contents in relation to host-feeding in some parasitic Hymenoptera. Entomophaga 40:87-93.

Levine, L., and D. J. Sullivan. 1983. Intraspecific tertiary parasitoidism in Asaphes lucens (Hymenoptera: Pteromalidae), an aphid hyperparasitoid. Canadian Entomologist 115:1653-1658.

Mackauer, M., and W. Völkl. 1993. Regulation of aphid populations by aphidiid wasps: does parasitoid foraging behaviour or hyperparasitism limit impact? Oecologia 94:339-350.

Matejko, I., and D. J. Sullivan S.J. 1980. Bionomics and behavior of Alloxysta megourae , and aphid hyperparasitoid (Hymenoptera: Alloxystidae). Journal of the New York Entomological Society 87:275-282.

Mayhew, P. J., and T. M. Blackburn. 1999. Does development mode organize life-history traits in the parasitoid Hymenoptera? Journal of Animal Ecology 68:906-916.

Mertins, J. W. 1985. Hyperparasitoids from pea aphid mummies, Acyrthosiphon pisum (Homoptera: Aphididae), in North America. Annals of the Entomological Society of America 78:186-195.

Pare, F., C. Cloutier, L. Huot, and J. N. McNeil. 1979. Description of egg and larval stages of Aphidius nigripes . Annals of the Entomological Society of America 72:620-626.

Quicke, D. L. J. 1997. Parasitic wasps. Chapman and Hall.

Radcliffe, E. B. 1982. Insect pests of potato. Annual Review of Entomology 27:173-204.

Roff, D. A. 1992. The evolution of life histories. Chapman and Hall, New York.

SAS Institute. 1999. SAS/STAT User's guide, version 8. SAS Institute, Cary, North Carolina.

Schooler, S. S., A. R. Ives, and J. Harmon. 1996. Hyperparasitoid aggregation in response to variation in Aphidius ervi host density at three spatial scales. Ecological Entomology 21:249-258.

Shands, W. A., G. W. Simpson, C. F. W. Muesebeck, and H. E. Wave. 1965. Parasites of potato-infesting aphids in northeastern Maine. Maine Agricultural Experimental Station Technical Bulletin T-19.

Singh, R., and P. N. Srivastava. 1987. Bionomics of Alloxysta pleuralis , a cynipoid hyperparasitoid of an aphidiid parasitoid Trioxys indicus . Entomologia Experimentalis et Applicata 43:11-15.

Stary, P. 1988. Natural enemies. Pages 171-188 in A. K. Minks and P. Harrewijn, editors. Aphids. Their Biology, Natural Enemies and Control. World Crop Pests, volume 2B. Elsevier, Amsterdam.

Stearns, S. C. 1992. The evolution of life histories. Oxford University Press, Oxford.

Strand, M. R. 2000. Developmental traits and life history evolution in parasitoids. Pages 139-162 in M. E. Hochberg and A. R. Ives, editors. Parasitoid population biology. Princeton University Press, Princeton.

Sullivan, D. J. 1987. Insect hyperparasitism. Annual Review of Entomology 32:49-70.

Sullivan S.J., D. J., and R. van den Bosch. 1971. Field ecology of the primary parasites and hyperparasites of the potato aphid, Macrosiphum euphorbiae , in the East San Fransisco Bay area. Annals of the Entomological Society of America 64:389-394.

Sullivan, D. J., and W. Völkl. 1999. Hyperparasitism: Multitrophic ecology and behavior. Annual Review of Entomology 44:291-315.

Völkl, W., and T. Barczak. 1990. Habitat selection by the aphid hyperparasitoid Aphidencyrtus aphidivorus Mayr (Hyemnoptera: Encyrtidae). Polskie Pismo Entomologiczne 60:129-135.

Völkl, W., and P. Kranz. 1995. Nocturnal activity and resource utilization in the aphid hyperparasitoid Dendrocerus carpenteri . Ecological Entomology 20:293-297.

Walker, G. P., P. J. Cameron. 1981. The biology of Dendrocerus carpenteri (Hymenoptera: Ceraphronidae), a parasite of Aphidius species, and field observations of Dendrocerus species as hyperparasites of Acythosiphon species. New Zealand Journal of Zoology 8: 531-538.

Walker, G. P., L. R. Nault, and D. E. Simonet. 1984. Natural mortality factors acting on potato aphid ( Macrosiphum euphorbiae ) populations in processing-tomato fields in Ohio. Environmental Entomology 13:724-732.

Table 2-1 Selected aphid hyperparasitoids and their taxonomic position (family within Hymenoptera), parasitic relationships of the larva (endo/ectoparasitoid, development mode and host stage), and host range. Host range is indicated as the number of aphid parasitoid genera that are parasitised.

Species	Family	Endo/ecto parasitoid	Development mode	Host stage	Host range

Asaphes suspensus	Pteromalidae	Ecto-	Idiobiont	(Pre-)pupa in aphid mummy	5
Dendrocerus carpenteri	Megaspilidae	Ecto-	Idiobiont	(Pre-)pupa in aphid mummy	4
Alloxysta victrix	Charipidae	Endo-	Koinobiont	Larva in live aphid	1
Syrphophagus aphidivorus	Encyrtidae	Endo-	Koinobiont	Larva in live aphid or (pre-)pupa in aphid mummy	4

Table 2-2 Life history parameters and intrinsic rates of increase for four aphid hyperparasitoids (means with standard error between brackets) reared on the Aphidius nigripes , Macrosiphum euphorbiae , potato system. Within rows, means followed by the same letter a are not significantly different based on Bonferroni–adjusted PLSD test. For differences between species, lower case letters are used for females and upper case letters for males (ANOVA/ANCOVA see text for F and p-values). N is the number of individuals tested.

	Alloxysta victrix		Syrphophagus aphidivorus		Dendrocerus carpenteri		Asaphes suspensus
	♀	♂	♀	♂	♀	♂	♀	♂
Developmental strategy	Endo parasitic Koinobiont		Endo parasitic Koinobiont		Ecto parasitic Idiobiont		Ecto parasitic Idiobiont

Development time (days)	19.5 (0.7) c	17.8 (0.8) B	17.7 (0.1) b	17.4 (0.1) B	15.7 (0.2) a	14.6 (0.03) A	21.4 (0.6) d		20.5 (0.3) C
(egg-adult)	N=14	N=4	N=172	N=96	N=136	N=129	N=21		N=19

Larval mortality (%)	3.0 (0.8) a		4.5 (0.7) a		5.2 (1.2) a		6.6 (1.6) a
Pupal mortality (%)	5.0 (1.7) a		3.0 (0.4) a		3.5 (1.2) a		2.0 (0.2) a
Unemerged adult mortality (%)	12.4 (2.3) b		1.4 (0.2) a		2.5 (0.5) a		2.5 (0.7) a
Total immature mort. %	20.4 (3.4) b		8.9 (0.7) a		11.3 (2.4) a		11.1 (1.8) a
	N=11		N=12		N=13		N=15

Longevity (days)	43.5 (1.5) b	41.7 (1.1) B	37.4 (3.2) a	22.2 (1.8) A	39.6 (1.6) ab	25.7 (1.2) A	70.3 (2.0) c	47.5 (2.2) C
	N=69	N=62	N=50	N=33	N=56	N=30	N=100	N=51

Realised fecundity	86 (12.3) a		577 (52.5) b		154 (11.6) a		834 (138) c
Potential fecundity	108 (16.0) a		629 (56.7) b		175 (12.5) a		924 (139) c
Pre-ovipos. period (days)	2.1 (0.5) b		0 (0) a		0.1 (0.1) a		1.1 (0.1) b
Oviposition period (days)	32 (2.7) a		26 (1.2) a		24 (2.2) a		59 (7.0) b
Post-ovipos. period (days)	2.4 (0.6) a		17.9 (4.5) ab		19.1 (3.5) b		16.4 (7.5) a b
	N=9		N=11		N=15		N=9

Sex ratio (% males)	25.0 a		70.2 c		39.5 b		76.7 c
	N=11		N=11		N=13		N=7

Dry weight (μg)	102.2 (2.4) bc	91.4 (2.1) B	55.7 (1.1) a	44.9 (1.7) A	108.9 (3.1) c	96.4 (3.5) B	98.5 (3.7) b	87.0 (3.9) B
Head width (μm)	426.1 (2.4) a	402.9 (5.2) B	423.5 (1.5) a	389.5 (5.5) A	549.5 (9.1) c	566.9 (3.2) D	500.9 (5.7) b	481.6 (4.9) C
Wing length (μm)	1770.3 (12.8) d	1669.2 (11.8) D	1091.9 (3.2) a	1032.6 (7.7) A	1440.0 (23.5) b	1405.5 (7.9) B	1570.0 (21.3) c	1455.8 (17.8) C
	N=33	N=42	N=85	N=19	N=65	N=29	N=22	N=19

Intrinsic rate of increase (rm) (d-1)	0.1180		0.2194		0.1712		0.1844

Table 2-3 Comparison of four aphid hyperparasitoid genera, relative to their conditions as predicted by the koinobiont/idiobiont dichotomous hypothesis (Quicke, 1997; Mayhew and Blackburn, 1999). Matching of Alloxysta and Syrphophagous is to predicted koinobiont traits, and Asaphes and Dendrocerus matching is to idiobionts.

Predicted for koinobionts	Alloxysta	Syrphophagus	Asaphes	Dendrocerus	Predicted for idiobionts
Specialist	Yes	No	Yes	Yes	Generalist
Host exposed	Yes	Yes	Yes	Yes	Concealed host
Host stage attacked smaller than wasp	Yes	Yes	No/yes	No/yes	Host stage attacked larger than wasp
No or temporary paralysis	Yes	Yes	Yes	Yes	Paralysis
Slow or delayed development4	Yes	Yes	No	Yes	Rapid development4
Small eggs4	Yes1	No1	Yes1	Yes1	Large eggs4
High oviposition rate4	No	Intermediate	Yes	Yes	Low oviposition rate4
High fecundity4	No	Yes	No	Yes	Low fecundity3
Pro-ovigeny	No	No	Yes	Yes	Synovigeny
Host feeding uncommon	?	No	Yes	No	Host feeding common
No oösorption	?	?	Yes2	No2	Oösorption
Short adult lifespan4	No	No	Yes	Yes	Long adult lifespan3
Large adult size	Yes	No	No	No	Small adult size
Less or no sexual dimorphism	No?	No?	Yes	Yes	Sexual dimorphism
Do not choose sex to match host size	?	?	Yes3	Yes	Choose sex to match host size

1Egg size for Alloxysta brevis, Asaphes vulgaris and D. carpenteri (Haviland, 1920, 1922, Christiansen-Weniger, 1992; Mackauer and Völkl, 1993) , Syrphophagus inquisitor (Griswold, 1929).

2Data for Asaphes vulgaris and D. carpenteri (LeRalec, 1995).

3Data for Asaphes vulgaris (Sullivan and Völkl, 1999).

4Demonstrated by Mayhew and Blackburn (1999) for parasitoids.

Figure 2-1Realised fecundity (mean ± SE; Alloxysta victrix n=9; Asaphes suspensus n=7; Dendrocerus carpenteri n=13; Syrphophagus aphidivorus n=11) as affected by female age of four aphid hyperparasitoids reared on the Aphidius nigripes , Macrosiphum euphorbiae , potato system

Figure 2-2 Sex ratio of progeny (mean ± SE; Alloxysta victrix n=11; Asaphes suspensus n=7; Dendrocerus carpenteri n=13; Syrphophagus aphidivorus n=11) as affected by female age of four aphid hyperparasitoids reared on the Aphidius nigripes , Macrosiphum euphorbiae , potato system.


Chapter 1. General Introduction		Chapter 3. Preference and performance of the hyperparasitoid Syrphophagus aphidivorus (Hymenoptera: Encyrtidae): Fitness consequences of selecting hosts in live aphid vs. aphid mummy