BEHAVIOUR, 2008, 75, 1423e1430 doi:10.1016/j.anbehav.2007.09.016

Available online at www.sciencedirect.com

Unusual mechanisms involved in learning of oviposition-induced host plant odours in an egg parasitoid?

ROLAND SCHRO¨ DER*,LARISSAWURM*, MARTTI VARAMA†, TORSTEN MEINERS* & MONIKA HILKER* *Institute of Biology, Freie Universita¨t Berlin yFinnish Forest Research Institute, Vantaa Research Centre

(Received 30 April 2007; initial acceptance 25 June 2007; final acceptance 29 September 2007; published online 19 November 2007; MS. number: 9378)

Parasitoids of herbivorous larvae are well known to be able to associatively learn a response to feeding- induced plant volatiles. However, little is known of the abilities of egg parasitoids to learn a response to herbivore oviposition-induced plant volatiles (OVIV). The eulophid wasp Chrysonotomyia ruforum is known to be attracted by OVIV from Scots pine released 3 days after host egg deposition. This study aimed to elucidate (1) the impact of the nutritional state of the parasitoid on its response to OVIV and (2) the effect of learning and response to the temporal change of OVIV. Supply of food as a reward during associative learning trials inevitably influences the nutritional state and thus leads to an experimental constraint. Nev- ertheless, the results indicate that the parasitoids have to associate OVIV with host presence and that they need to be supplied with food to learn and to respond to OVIV. Furthermore, the parasitoids’ positive response to OVIV from pine twigs laden with eggs for 3 days was independent of the induction time of pine twigs experienced. However, a positive response was given only to OVIV from pine twigs with 3-day-old eggs, while OVIV from twigs with other induction times (1, 2, 4 days) were not attractive. Thus, our findings indicate an unusual learning process in this specialized egg parasitoid. We suggest that the parasitoid is using a learning strategy by which common information is filtered out of similar odour blends (here: 1- to 4-day-long induced pine twigs), but a behavioural response to this learned infor- mation becomes evident only in a specific odour context (here: 3-day-long induced twigs).

Ó 2007 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Keywords: behavioural plasticity; Chrysonotomyia ruforum; Diprion pini; experience; nutrition; odour learning; parasitoid; Pinus sylvestris

In the past three decades associative learning of plant learning of complex plant volatile blends are not well odours has been studied in several species (Gould elucidated yet. 1993; Papaj & Lewis 1993; Smid 2006). It became evident Parasitic wasps attacking herbivorous are known that many insects are able to associate specific plant to use both plant and herbivore odour to locate their hosts odours as a conditioned stimulus with a rewarding uncon- (Vet et al. 1995; Quicke 1997; Vinson 1998; Steidle & van ditioned stimulus such as food and/or host resources (e.g. Loon 2003; Takabayashi et al. 2006). The volatiles released Daly & Smith 2000; Laloi et al. 2000; Meiners et al. 2003). from plants and herbivores may vary due to numerous However, the exact conditions necessary for associative abiotic and biotic factors, thus exposing the parasitoids to a highly complex and variable odourous environment (Hilker & McNeil, in press). Parasitoid responses to phenotypic changes of plant volatiles have been studied Correspondence: M. Hilker, Institute of Biology, Freie Universita¨t Berlin, Haderslebener Str. 9, D-12163 Berlin, Germany (email: hilker@zedat. intensively with respect to changes induced by herbivore fu-berlin.de). R. Schro¨der is now at the Department of Zoology, Univer- attack or herbivore egg deposition (Turlings & Benrey sity of Hawaii, 2538 McCarthy Mall, Edmondson Hall 259, 96822 1998; Kessler & Baldwin 2001; Hilker & Meiners 2002, Honolulu, HI, U.S.A. M. Varama is at Finnish Forest Research Institute, 2006; Turlings et al. 2002; Mumm & Hilker 2006). The Vantaa Research Centre, PO Box 18, FIN-01301 Vantaa, Finland. changes of plant odour may be specific for the herbivore 1423 0003e3472/08/$34.00/0 Ó 2007 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. 1424 ANIMAL BEHAVIOUR, 75,4

species and herbivore phenotype (De Moraes et al. 1998; whether the na€ıve parasitoids tested so far did not respond Meiners et al. 2000; Turlings et al. 2002; Mumm et al. to OVIV simply because they were hungry and not fed 2005), thus providing information on the most suitable prior to the experience phase and bioassay. host. To cope with high variability of plant volatiles, Second, we investigated the parameters (rewards) that a parasitoid should show appropriate phenotypic plastic- might be important during the experience phase and thus ity to successfully locate hosts. might be necessary for associative learning by C. ruforum. Parasitoid responses to plant odour are known to change We hypothesized that parasitoids fed honey prior to the due to experience (Vet et al. 1995) but also due to age experience phase need to associate during an experience (Kester & Barbosa 1991; Bertschy et al. 1997), egg load phase OVIV (conditioned stimulus) with (1) food as (Vet & Dicke 1992) and nutritional status (Lewis & Takasu reward (unconditioned stimulus), (2) host eggs but no 1990; Jacob & Evans 2001). The ability to learn by experi- food as reward or (3) both food and host eggs as rewards ence is widespread among parasitoids (Turlings et al. 1993; (Gould 1993). Steidle & van Loon 2003; Smid 2006) and has been found Third, we aimed to elucidate whether the positive to significantly improve host and food finding (Lewis & response of the parasitoids to OVIV from pine twigs laden Martin 1990; Papaj & Vet 1990; Steidle 1998). Many adult with 3-day-old eggs was affected by experience of OVIV parasitoids are able to learn associatively to respond to cues from pine twigs with induction times other than those they have experienced in the presence of hosts or food tested so far (1 day; Hilker et al. 2002). Additionally, we ex- (Turlings et al. 1993; Vet et al. 1995; Smid 2006). amined the attractiveness of odours that are emitted from Learning abilities may be essential for parasitoids attack- pine twigs after different time points (1e4 days) of sawfly ing many host species or hosts with a wide range of host egg deposition. It is well known that herbivore-induced plants (Vet & Dicke 1992), but they may even support volatiles change over time (Loughrin et al. 1994; Turlings foraging by specialized parasitoids that also have a genetic et al. 1998). Thus, the response of the egg parasitoid to predisposition to specific cues (Steidle & van Loon 2003). odour of pine twigs laden with eggs may change with If specific host or food cues are available only in low changing induction times. amounts or only during a short time period, a learned response to other more general cues also associated with the host or food might serve as ‘back-up’ trait for special- METHODS ized parasitoids to intensify or prolong successful foraging. For example, the response of the egg parasitoid Chrysono- Plants and Insects tomyia ruforum to pine volatiles induced by host egg deposition may be considered such a back-up trait (Hilker Large branches of Pinus sylvestris L. (Pinales, Pinaceae) et al. 2002). This eulophid wasp is specialized on diprionid were detached from 15- to 30-year-old trees in a forest sawflies feeding on pine (Pschorn-Walcher & Eichhorn near Berlin, Germany and brought into the laboratory 1973; Eichhorn & Pschorn-Walcher 1976; Pschorn- where the stems were cleaned and sterilized according to Walcher 1982). It is known to respond innately to sex the method of Moore & Clark (1968). For all experiments, pheromones of the most abundant host species, that is, small twigs (about 20 cm length) were cut from these the sawflies Neodiprion sertifer and Diprion pini (Hilker branches and treated as described below. et al. 2000). In addition to the innate response to host The sawfly D. pini L. (, Diprionidae) was cues, this egg parasitoid is attracted to oviposition- reared continuously in the laboratory on cut pine induced pine volatiles (OVIV) when these have been expe- branches (about 60 cm length) as described by Bombosch rienced before. Odour of pine laden with 3-day-old eggs is & Ramakers (1976) and Eichhorn (1976) at 20 1C, 65% attractive (Hilker et al. 2002). Whereas the presence of sex RH and 18:6 h light:dark (LD) cycles. Adult females lay up pheromones might indicate future egg laying or freshly to 20 eggs per needle (about 100e150 eggs per female). laid host eggs to C. ruforum, the OVIV indicate that eggs The larvae of this species may heavily damage pine forests. have been laid 3 days before. Thus, the ability to detect Cocoons overwinter mainly in the soil (Pschorn-Walcher OVIV might serve as a back-up trait for C. ruforum to locate 1982; Brauns 1991). host eggs even when sex pheromones have evaporated. The egg parasitoid C. ruforum (Krausse; Hymenoptera, Na€ıve and hungry C. ruforum were found not to respond ) was obtained from parasitized host eggs col- to oviposition-induced pine volatiles (Mumm et al. 2005). lected in the field in southern Finland. This parasitoid spe- So far, this egg parasitoid was shown to respond to odours cies may considerably limit population densities of D. pini emitted from pine twigs after 3 days of sawfly egg deposi- (Pschorn-Walcher 1982). In addition to D. pini, eggs of tion only when it has experienced the plantehost complex a few other diprionid species feeding on pine are attacked with additional availability of honey as food source. The also by C. ruforum (Eichhorn & Pschorn-Walcher 1976). plantehost complex offered during experience consisted Pine needles with parasitized eggs were brought into the of pine twigs that emitted odours after 1 day of sawfly laboratory and kept in petri dishes at 5C. To initiate egg deposition (Hilker et al. 2002; Mumm et al. 2005). parasitoid emergence, needles with parasitized eggs were The study presented here aimed to elucidate the para- transferred to 20 1C, 65% RH and 18:6 h LD. Emerging meters that affect learning and thus response of C. ruforum adults were collected daily, transferred into a 20-ml vial to OVIV released from pine twigs. and kept at 10 1C, 65% RH and 18:6 h LD until they First, we studied the importance of the nutritional state were used for conditioning. The time between emergence of the egg parasitoid and investigated the question of and conditioning is referred to here as ‘storage phase’. Due SCHRO¨ DER ET AL.: EGG PARASITOID LEARNING 1425 to an irregular hatching rate it was not possible to perform a cage (3.5 litres) in the evening with a single twig in tap bioassays with parasitoids of a specific age. Thus, 2- to water to allow mating and egg deposition (abiotic condi- 14-day-old female parasitoids were tested (see also Hilker tions: 20 1C, 65% RH and 18:6 h LD). The next morning et al. 2002; Mumm et al. 2005). To prevent possible age when at least four egg masses had been laid was referred to effects (e.g. Kester & Barbosa 1991; Takasu & Lewis as day 0. At day 0, sawflies were removed from the twigs. 1996), we randomized the distribution of age classes of From this time on, the pine twigs were kept at the same abi- the wasps tested within each bioassay. otic conditions for an additional 1e4 days. Thus, pine twigs that emit odours after different time points of sawfly egg deposition could be used for conditioning and testing. Bioassays: General Procedure and Statistics Twigs used for bioassays were removed from the tap water in which they were kept, tightly wrapped with parafilm The response of parasitoids with different nutritional at the cut end and introduced in the olfactometer test flask. states and different experiences of OVIV was tested in For conditioning, parasitoids were added to these twigs to a four-arm airflow olfactometer (Vet et al. 1983; Hilker give them experience with the plantehost complex (see et al. 2002). One field was supplied by an odour from below for detailed information). a twig kept in a test flask (250 ml) and the other three fields were supplied with charcoal-filtered control air. 1 The airflow was adjusted to 155 ml min . When starting General Conditioning Procedure the bioassay, a single parasitoid female was introduced into the arena. The time the parasitoid spent walking in Parasitoids were used after their storage phase (time each of the four odour fields was recorded over an observa- between emergence and conditioning) for a conditioning tion period of 600 s using the software The Observer 3.0 period of 1 day (‘conditioning or experience phase’; (Noldus, Wageningen, The Netherlands). Only data ob- Fig. 1). During an experience phase approximately 15 tained by observation of active parasitoids walking at least female and 15 male parasitoids were placed into a cage 300 s of the bioassay period (600 s) were used for statistical (3.5 litres) with a single twig in tap water for 1 day to allow analysis. Data were analysed by Wilcoxon signed-ranks mating and experience of the plantehost complex at tests by comparing parasitoid walking times in the test 20 1C, 65% RH and 18:6 h LD. Thereafter, female par- field and in the opposite control field. The two other asitoids were separated from the males and kept an addi- control fields located between the test and the opposite tional period of 1 day (‘lag phase’) without any contact control field were considered buffer fields (Ninkovic to the plantehost complex at conditions described for et al. 2001). The number of parasitoids and odour sources storage (see above). After the lag phase the bioassays tested are given in Tables 1, 2. The analyses were were conducted. performed using Statistica 4.5 scientific software (Statsoft, Hamburg, Germany). Na€ıve Parasitoids: Nutritional State

Treatment of Pine Twigs To test whether feeding parasitoids with honey is sufficient for even na€ıve parasitoids to respond to OVIV, Small twigs (about 20 cm length) were cut and placed in we fed the parasitoids with honey water during storage, a vial with tap water. To obtain oviposition-induced pine conditioning and lag phase. Parasitoids had no contact twigs, two female and two male sawflies were placed into with the plantehost complex during the conditioning

Table 1. Effects of nutrition and experience for egg parasitoid learning of oviposition-induced pine volatiles Experience Duration of walking (s) Storage Lag Host egg Food OVIV Food contact Food Test field Opposite field N parasitoid N plant PT

OVIV na€ıve a1 þ 0 þ 0 þ 96 (13e211) 100 (25e151) 31 8 0.570 219 Associative learning b1 þþþ0 þ 156 (90e255) 108 (40e169) 31 5 0.147 174 b2 0 þ 0 þ 0 94 (41e172) 111 (59e148) 31 8 0.799 235 b3 þþ0 þ 0 122 (60e301) 62 (14e222) 30 5 0.644 210 b4 þþþþþ205 (80e291) 47 (7e143) 31 9 0.003 98

Response of na€ıve (a1) and conditioned (b1 to b4) female Chrysonotomyia ruforum to odours emitted from pine twigs after 3 days of sawfly egg deposition (test field) in a four-field olfactometer (opposite field of test field ¼ control field with clean air). Columns ‘Storage’ and ‘Lag’ indicate whether the parasitoids were supplied with food (þ) or not (0) during their storage and lag phases. Column ‘Experience’ indicates what parasitoids experienced prior to olfactometer bioassays: odours emitted from pine twigs after 1 day of sawfly egg deposition (OVIV) with different rewards of food and/or host eggs. Medians and interquartile range (parentheses) for the time the parasitoid female spent walking in the olfactometer fields over an observation period of 600 s are given. P value is evaluated by Wilcoxon signed-ranks test and T value of the test statistic is given. 1426 ANIMAL BEHAVIOUR, 75,4

Table 2. Odour from pine after different induction times Duration of walking (s)

Odour in experience (days) Odour in bioassay (days) Test field Opposite field N parasitoid N plant PT

Conditioned and tested pine odours are the same A1 1 1 155 (82e232) 89 (23e194) 35 8 0.145 226 A2 2 2 123 (65e190) 132 (73e197) 35 5 0.866 182 A3 3 3 225 (93e343) 52 (11e119) 38 10 <0.001 123 A4 4 4 121 (41e209) 99 (59e149) 30 4 0.530 202 Conditioned and tested pine odours are different B1 1 3 362 (294e413) 33 (10e62) 27 7 <0.001 0 B2 2 3 160 (70e251) 69 (7e130) 46 10 0.011 308

Learning and response of the egg parasitoid C. ruforum. Response of fed female parasitoids to odours emitted from pine twigs (odour in bioassay, in test field) in a four-arm olfactometer (opposite field of test field ¼ control field with clean air). Prior to olfactometer bioassays par- asitoids were fed and enabled to associate OVIV from pine twigs with different induction periods (1, 2, 3 and 4 days) (¼ odour in experience) with host eggs and food. Odours from pine twigs after 1e4 days of sawfly egg deposition were used for the bioassays (odour in bioassay). Experienced and tested pine odours were the same (A1eA4); experienced and tested odours were different (B1, B2). Medians and interquartile ranges (parentheses) for the time the parasitoid female spent walking over an observation period of 600 s are given. P value is evaluated by Wilcoxon signed-ranks test and T value of the test statistic is given.

phase. For the bioassay, we used odours emitted from pine the lag phase (i.e. after conditioning; Table 1, b3). These twigs 3 days after sawfly egg deposition (Table 1, a1). bioassays allowed testing the effect of nutrition also in ex- perienced parasitoids. We did not feed parasitoids during the lag phase to prevent a delayed association of OVIV Experienced Parasitoids: Associative Learning with food because hungry parasitoids that were trained and fed afterwards may respond to trained odours in a Associative learning abilities of the egg parasitoid were way similar to that of trained and fed parasitoids (Lewis & investigated using a plantehost complex consisting of pine Takasu 1990). twigs laden with sawfly eggs for 1 day for the conditioning The question of whether parasitoids are able to associate phase. For the bioassays, the parasitoid’s response to OVIV both food and host egg contact with OVIV was investi- from 3-day-long induced pine twigs were used, that is, from gated by giving parasitoids access to honey water and pine twigs laden with sawfly eggs for 3 days (see below for leaving the egg-laden pine twig without gauze to allow other induction times). To test the different hypotheses on host egg contact and parasitism during the conditioning associative learning of the parasitoid (see Introduction), we phase. Additionally, parasitoids were fed with honey water used the following setups (Table 1). during storage and lag phases (Table 1, b4). The parasitoid’s ability to associate OVIV with food only as reward was tested by giving parasitoids access to honey water during the conditioning phase, while Odour from Pine after Different Induction e contact with the plant host complex was prevented by Times: Learning and Response wrapping the egg-laden twig with a gauze (mesh width: 50 mm) that allowed release of OVIV. Additionally, parasit- To investigate whether the induction time of pine twigs oids were fed with honey water during storage and lag during the conditioning process affects learning and thus phases (Table 1, b1). the parasitoid’s response to OVIV, we used odours that are The parasitoid’s ability to associate OVIV with host eggs emitted from pine twigs after different time points (1e4 only as reward was studied by leaving the egg-laden pine days) of sawfly egg deposition during conditioning. Access twig without gauze to allow host egg contact and parasit- to honey water was given, and the egg-laden pine twig ism during the conditioning phase. However, no access to was left without gauze to allow host egg contact and honey water was given for the parasitoids during this parasitism. Parasitoids were fed with honey water during phase. Two sets of parasitoids were tested: parasitoids that storage and lag phases. were deprived of honey during the storage and lag phases First, the parasitoid’s response to OVIV from pine twigs (Table 1, b2) and parasitoids that were fed honey during with an induction time (1, 2, 3 and 4 days) that had been the storage phase (i.e. prior to conditioning) but not in experienced in the conditioning phase was tested; that is,

∗ Emergence Storage Experience Lag phase Bioassay phase phase (1 day) (2–14 days) (conditioning) (1 day) Figure 1. General temporal scheme for handling parasitoids. SCHRO¨ DER ET AL.: EGG PARASITOID LEARNING 1427 the parasitoids experienced OVIV from pine twigs with 1 pine volatiles only when being fed and given the chance to (2, 3 and 4)-day-old eggs and their response to OVIV from experience OVIV in the presence of both host eggs and pine twigs with 1 (2, 3 and 4)-day-old eggs was tested food. Thus, the previous finding that na€ıve and hungry (Table 2,A1eA4). parasitoids do not respond to OVIV (Hilker et al. 2002; Furthermore, to study whether the response of the egg Mumm et al. 2005) was due not to the absence of food parasitoid to OVIV released from pine twigs with a but to the lack of a chance to associate OVIV with a reward. 3-day-long induction time (known to be attractive; Hilker Our present results clearly show that the egg parasitoid et al. 2002) is affected when OVIV released from pine needs to have the chance to associate OVIV with other twigs with other induction times have been associated cues from the plantehost complex to enable a positive re- with a reward, we tested the response of parasitoids to sponse to OVIV. What are the mechanisms involved in as- OVIV from pine twigs with 3-day-old eggs after parasitoids sociative learning of OVIV in this specialized egg experienced OVIV from pine twigs with 1- and 2-day-old parasitoid? Which parameters need to be associated eggs. Even though parasitoids experienced with OVIV with OVIV? from pine twigs with 1-day-old eggs are known to be The hypothesis that C. ruforum associates OVIV (condi- attracted by odour from pine twigs with 3-day-old eggs tioned stimulus) with food (unconditioned stimulus) (Hilker et al. 2002), we repeated this bioassay to check must be rejected because parasitoids that were exposed our previous results (Table 2, B1 and B2). to OVIV and received food, but were deprived of host eggs, did not respond to OVIV (Table 1, b1). Neither did unfed nor fed parasitoids respond to OVIV when they RESULTS had been exposed to OVIV and host eggs, but not to food, during conditioning (Table 1, b2 and b3). These € Naıve Parasitoids: Nutritional State results suggest the rejection of the hypothesis that a posi- Fed parasitoids without any experiences of the plante tive response of parasitoids to OVIV requires association host complex were not attracted to oviposition-induced of OVIV with host eggs only. Furthermore, one might pine volatiles (Table 1, a1). conclude from these results that the nutritional state did not affect the parasitoid’s association of OVIV with host eggs because feeding during the storage phase did not in- Experienced Parasitoids: Associative Learning fluence the results. However, the rejection of this hypoth- esis and conclusions on the nutritional state need to take Parasitoids that were fed prior to conditioning and were into account the following: the nutritional state of the allowed to associate OVIV with food, but not with host parasitoids might have been determined not only by eggs, during the conditioning phase were not attracted to (non)feeding in the storage phase but also by (non)supply OVIV (Table 1, b1). Unfed parasitoids that were allowed to with food during the conditioning and storage phases. associate OVIV with host eggs, but not with food, during Many adult parasitoids need to feed on sugar sources to the conditioning phase were not attracted to OVIV (Table cover their energetic needs (reviewed by Wa¨ckers et al. 1, b2). This result did not change when parasitoids were 2005). Some parasitoids are known to need a ‘sweet sup- fed prior to the conditioning phase (Table 1, b3). Only ply’ at least once or twice a day (Siekmann et al. 2001; those parasitoids that were fed prior to conditioning and Azzouz et al. 2004). When C. ruforum was supplied with were allowed to associate OVIV with both food and host food during the storage phase, depriving of food during eggs were significantly attracted to OVIV (Table 1, b4). the conditioning phase (1 day) might have led to hungry individuals (Table 1, b2) and thus the nutritional state might have negatively influenced the learning process. Odour from Pine after Different Induction Furthermore, as argued above (Methods), in those bioas- Times: Learning and Response says where C. ruforum was deprived of food in the condi- tioning phase (Table 1, b1 and b2), no food was given also Parasitoids that were fed prior to conditioning and in the lag phase to prevent ‘postconditioning-phase- experienced food and odours emitted from pine twigs association’ with food (Lewis & Takasu 1990). However, laden with 1-, 2- or 4-day-old sawfly eggs were not depriving of food during the lag phase might lead to attracted to the pine odours they had experienced (Table hungry parasitoids that have a lowered response ability 2,A1eA4). However, parasitoids that had experienced to OVIV. odour from pine twigs laden with 3-day-old eggs were With regard to the hypothesis that parasitoids fed prior attracted to the odour of these twigs in the olfactometer to conditioning and given the chance to associate OVIV bioassay. They were also attracted to OVIV from pine twigs with both food and host eggs as rewards show a positive with 3-day-old eggs when they had experienced odours response to OVIV, only satiated parasitoids that experi- that were emitted from pine twigs with 1- or 2-day-old enced both food and eggs along with OVIV were attracted sawfly eggs (Table 2, B1 and B2). to OVIV (Table 1, b4). However, again, while we could elucidate that nutrition and satiation per se without any DISCUSSION chance of OVIV experience are not sufficient to trigger a positive response to OVIV and that learning is necessary, Our results show that an egg parasitoid specialized on pine the role of the nutritional state for this learning procedure sawflies is able to respond positively to oviposition-induced remains ambiguous. Providing food as reward during 1428 ANIMAL BEHAVIOUR, 75,4

associative learning inevitably influences the nutritional different induction times (Smith 1993; Vet et al. 1995). state. It is an experimental constraint that supply with However, this is not the case because only odour from food during the conditioning phase could be sufficient 3-day-long induced twigs elicits a positive response. to satiate the parasitoids and trigger the association of Because none of the learning modes mentioned above OVIV with host eggs. Thus, based on our results we can explain our findings, we suggest the following cannot decide whether food is necessary during the condi- scenario: C. ruforum is able to learn a certain part of the tioning phase as an unconditioned stimulus (reward) or blend of OVIV that is present in odour from pine twigs necessary to provide the physiological conditions that with any induction time tested. This part of the blend enable them to associate OVIV with host eggs. Future may be specific for both the particular compounds per se studies of the parasitoid’s feeding periodicity could help and the quantitative ratios in which they occur. This puta- determine the normal time between meals. Such data tive part of the blend common to 1-, 2- and 3-day-long could provide some indirect hints to answer the question induced twigs is associated with a reward. To positively whether the parasitoids deprived of food in our bioassays respond to OVIV, this learned part of the blend of OVIV were hungry. needs to be perceived in a highly specific odour context, Food-associated and host-associated learning are orga- that is, the odour context provided by the blend released nized separately and expressed in accordance with the from pine twigs with a 3-day-long induction period. Ac- physiological state of the individual parasitoid (Lewis & cording to this scenario, the parasitoids respond positively Takasu 1990; Chan & Godfray 1993; Lewis et al. 1998; to the learned pattern of volatiles common to pine twigs Jacob & Evans 2001). The two types of learning involve with 1-, 2- and 3-day-long induction time only if this is different unconditioned stimuli (food and oviposition) perceived in admixture with other volatiles released and trigger different unconditioned responses (Wa¨ckers from 3-day-long induced pine twigs. Thus, the behaviou- et al. 2002; Olson et al. 2003). If energy reserves drop to ral response to a learned pattern of volatiles depends on low levels, parasitoids should search for food instead of the odourous context in which this pattern is perceived. hosts as suggested by optimal time allocation decisions Context-specific response to an odour has been shown by parasitoids (Siekmann et al. 2004; Tenhumberg et al. in several insects (Akers & Getz 1993; Smith 1998; Hosler 2006). Thus, we suggest that only C. ruforum that were & Smith 2000; Duchamp-Viret et al. 2003; Meiners et al. satiated during conditioning were able to associate OVIV 2003; Said et al. 2005; Hilker & McNeil, in press). For with host eggs which offer several putative rewarding example, a parasitoid of Drosophila larvae showed an en- stimuli that occur during ovipositor probing, oviposition, hanced response to yeast in the presence of a green leaf host feeding or just any stimulus with a high response volatile, if this green volatile had been experienced potential (Vet et al. 1995). Honeydew excreted by aphids before together with yeast odour (Vet & Groenewold and other sap-feeding homopterous insects represent 1990). While the odour blend of 3-day-long oviposi- a potential carbohydrate source for parasitoids (Elliott tion-induced pine twigs has been analysed chemically et al. 1987; Idoine & Ferro 1988; Lee et al. 2004). Because and is known to contain a complex blend of mono- aphids are common insect pests of pine (e.g. Veteli et al. and sesquiterpenes (Mumm et al. 2003), the odour 2006), this carbohydrate source is also available for blends of pine twigs with shorter induction times have C. ruforum in nature. not been studied so far. Therefore, future studies need Interestingly, the parasitoids responded only to those to address the question of which of the volatiles of the OVIV that were emitted after a 3-day-induction period, complex blend of 3-day-long induced twigs are learned regardless of which OVIV (released after a 1-, 2- or 3-day- and which are those providing the specific context. Fur- induction period) was experienced. These latter results thermore, the question on the benefit of C. ruforum to indicate that even when a discrepancy between the blend respond only to odour from 3-day-long induced pine of cues experienced and the blend of cues offered for twigs requires further experiments on the suitability of response can be expected, the experience with each type different egg ages for parasitism. of oviposition-induced twigs triggered a highly specific In conclusion, the egg parasitoid studied shows highly response (Table 2), that is, a positive response only to interesting abilities to learn and to respond to host- OVIV released from 3-day-long induced twigs. Because associated odour. The odour of an unattractive twig (e.g. C. ruforum needs to associate OVIV with a reward, nonas- 1- or 2-day-long induced twigs) might have been learned sociative learning processes such as sensitization and like a part of a puzzle. The picture (here: the signal for priming (e.g. Turlings et al. 1993) cannot explain these re- a plant with suitable host eggs) that elicits a behavioural sults. Neither can associative learning of the entire com- response becomes clear only when another part of the plex odour blend of 1-, 2- or 3-day-long induced twigs puzzle is present (here: odour from 3-day-long-induced be an explanation because associative learning sensu twigs). These learning and response abilities suggest some strictu means that C. ruforum would be able to respond kind of ‘aha-experience’ (Bu¨hler 1907) referred to here as positively to exactly the same blend that they encoun- experience of something that has been perceived before tered during their conditioning phase (Gould 1993; Smith but makes sense only when presented in a specific 1993; Turlings et al. 1993). Neither can a putative ability context. The parasitoid is able to learn an odour possibly of the parasitoids to generalize traits common to odour by association of a part of a volatile blend with a reward; from pine twigs with 1-, 2- or 3-day-induction periods however, we suggest that this learned odour information explain the findings because such a generalization would triggers a positive behavioural response only when lead to the same response to odour from pine twigs with perceived in an odourous context that is specific for SCHRO¨ DER ET AL.: EGG PARASITOID LEARNING 1429 a certain ecological background (here the time of infesta- Hilker, M. & McNeil, J. In press. Chemical and behavioural ecology tion of pine twigs with host eggs). Furthermore, from in insect parasitoids: how to behave optimally in a complex odour- our experiments on the importance of the parasitoid’s ous environment. In: Behavioural Ecology of Insect Parasitoids, From nutritional state we conclude that the parasitoid needs Theoretical Approaches to Field Applications (Ed. by E. Wajnberg, to be supplied with sugars to enable successful learning. C. Bernstein & J. van Alphen) Oxford:Blackwell. Whether this sweet supply is necessary to provide an Hilker, M. & Meiners, T. 2002. Induction of plant responses to ovi- position and feeding by herbivorous : a comparison. adequate physiological state or whether it is associated Entomologia Experimentalis et Applicata, 104, 181e192. with the odour as reward is difficult to answer because Hilker, M. & Meiners, T. 2006. Early herbivore alert: insect eggs in- of experimental constraints. duce plant defense. Journal of Chemical Ecology, 32, 1379e1397. Hilker, M., Bla¨ske, V., Kobs, C. & Dippel, C. 2000. Kairomonal Acknowledgments effects of sawfly sex pheromones on egg parasitoids. Journal of Chemical Ecology, 26, 2591e2601. We thank Ute Braun (Freie Universita¨t Berlin, Germany) Hilker, M., Kobs, C., Varama, M. & Schrank, K. 2002. Insect egg for rearing D. pini. We also thank Martin Lehmann and deposition induces Pinus sylvestris to attract egg parasitoids. Jour- Ute Trauer for their help in conducting several of the bio- nal of Experimental Biology, 205, 455e461. assays. This research was funded by the German National Hosler, J. S. & Smith, B. H. 2000. Blocking and the detection of Research Foundation (DFG-GRK 837/1-03). odor components in blends. Journal of Experimental Biology, 203, 2797e2806. Idoine, K. & Ferro, D. N. 1988. Aphid honeydew as a carbohydrate References source for Edovum puttleri (Hymenoptera: Eulophidae). Environ- mental Entomology, 17, 941e944. Akers, R. P. & Getz, W. M. 1993. Response of olfactory receptor Jacob, H. S. & Evans, E. W. 2001. Influence of food deprivation on neurons in honeybees to odorants and their binary-mixtures. foraging decisions of the parasitoid Bathyplectes curculionis (Hyme- Journal of Comparative Physiology A, 173, 169e185. noptera: Ichneumonidae). Annals of the Entomological Society of Azzouz, H., Giordanengo, P., Wa¨ckers, F. L. & Kaiser, L. 2004. America, 94, 605e611. Effects of feeding frequency and sugar concentration on behavior Kessler, A. & Baldwin, I. T. 2001. Defensive function of herbi- and longevity of the adult aphid parasitoid: Aphidius ervi (Haliday) vore-induced plant volatile emissions in nature. Science, 291, (Hymenoptera: Braconidae). Biological Control, 31, 445e452. 2141e2144. Bertschy, C., Turlings, T. C. J., Bellotti, A. C. & Dorn, S. 1997. Chem- Kester, K. M. & Barbosa, P. 1991. Postemergence learning in insect ically-mediated attraction of three parasitoid species to mealybug- parasitoid, Cotesia congregata (Say) (Hymenoptera: Braconidae). infested cassava leaves. Florida Entomologist, 80, 383e395. Journal of Insect Behavior, 4, 727e742. Bombosch, S. & Ramakers, P. M. J. 1976. Zur Dauerzucht von Laloi, D., Bailez, O., Blight, M. M., Roger, B., Pham-Delegue, Gilpinia hercyniae Htg. Zeitschrift fu¨r Pflanzenkrankheiten und M. H. & Wadhams, L. J. 2000. Recognition of complex odours Pflanzenschutz, 83,40e44. by restrained and free-flying honey bees, Apis mellifera. Journal of Brauns, A. 1991. Taschenbuch der Waldinsekten. Stuttgart: Gustav Chemical Ecology, 26, 2307e2319. Fischer-Verlag. Lee, J. C., Heimpel, G. E. & Leibee, G. L. 2004. Comparing floral Bu¨hler, K. 1907. Tatsachen und Probleme zu einer Psychologie der nectar and aphid honeydew diets on the longevity and nutrient Denkvorga¨nge. Archiv fu¨r die Gesamte Psychologie, 9, 297e365. levels of a parasitoid wasp. Entomologia Experimentalis et Applicata, e Chan, M. S. & Godfray, H. C. J. 1993. Host-feeding strategies of 111, 189 199. parasitoid wasps. Evolutionary Ecology, 7, 593e604. Lewis, W. J. & Martin, W. R. 1990. Semiochemicals for use with Daly, K. C. & Smith, B. H. 2000. Associative olfactory learning in parasitoids: status and future. Journal of Chemical Ecology, 16, e the moth Manduca sexta. Journal of Experimental Biology, 203, 3067 3089. 2025e2038. Lewis, W. J. & Takasu, K. 1990. Use of learned odours by a par- De Moraes, C. M., Lewis, W. J., Pare´, P. W., Alborn, H. T. & asitic wasp in accordance with host and food needs. Nature, e Tumlinson, J. H. 1998. Herbivore-infested plants selectively at- 348, 635 636. tract parasitoids. Nature, 393, 570e573. Lewis, W. J., Stapel, J. O., Cortesero, A. M. & Takasu, K. 1998. Duchamp-Viret, P., Duchamp, A. & Chaput, M. A. 2003. Single Understanding how parasitoids balance food and host needs: im- e olfactory sensory neurons simultaneously integrate the compo- portance to biological control. Biological Control, 11, 175 183. nents of an odour mixture. European Journal of Neuroscience, 18, Loughrin, J. H., Manukian, A., Heath, R. R., Turlings, T. C. J. & 2690e2696. Tumlinson, J. H. 1994. Diurnal cycle of emission of induced vola- Eichhorn, O. 1976. Dauerzucht von Diprion pini L. (Hym.: Diprioni- tile terpenoids herbivore-injured cotton plants. Proceedings of the e dae) im Laboratorium unter Beru¨cksichtigung der Fotoperiode. National Academy of Sciences, U.S.A. 91, 11836 11840. Anzeiger fu¨r Scha¨dlingskunde, 49,38e41. Meiners, T., Westerhaus, C. & Hilker, M. 2000. Specificity of chem- Eichhorn, O. & Pschorn-Walcher, H. 1976. Studies on the biology ical cues used by a specialist egg parasitoid during host location. e and ecology of the egg-parasites (Hym. Chalcidoidea) of the pine Entomologia Experimentalis et Applicata, 95, 151 159. sawfly Diprion pini (L.) (Hym. Diprionidae) in Central Europe. Meiners, T., Wa¨ckers, F. L. & Lewis, W. J. 2003. Associative learn- Zeitschrift fu¨r Angewandte Entomologie, 80, 355e381. ing of complex odours in parasitoid host location. Chemical Senses, e Elliott, N. C., Simmons, G. A. & Sapio, F. J. 1987. Honeydew and 28, 231 236. wildflowers as food for the parasites Glypta fumiferanae (Hymenop- Moore, G. F. & Clark, E. W. 1968. Suppressing microorganisms and tera: Ichneumonidae) and Apanteles fumiferanae (Hymenoptera: maintaining turgidity in coniferous foliage used to rear insects in Braconidae). Journal of the Kansas Entomological Society, 60,25e29. the laboratory. Journal of Economic Entomology, 61, 1030e1031. Gould, J. L. 1993. Ethological and comparative perspectives on honey Mumm, R. & Hilker, M. 2006. Direct and indirect chemical de- bee learning. In: Insect Learning (Ed. by D. R. Papaj & A. C. Lewis), fence of pine against folivorous insects. Trends in Plant Science, pp. 18e51. New York: Chapman & Hall. 11, 351e358. 1430 ANIMAL BEHAVIOUR, 75,4

Mumm, R., Schrank, K., Wegener, R., Schulz, S. & Hilker, M. Takabayashi, J., Sabelis, M. W., Janssen, A., Shiojiri, K. & van 2003. Chemical analysis of volatiles emitted by Pinus sylvestris after Wijk, M. 2006. Can plants betray the presence of multiple her- induction by insect oviposition. Journal of Chemical Ecology, 29, bivore species to predators and parasitoids? The role of learning 1235e1252. in phytochemical information networks. Ecological Research, 21, Mumm, R., Tiemann, T., Varama, M. & Hilker, M. 2005. Choosy 3e8. egg parasitoids: specificity of oviposition-induced pine volatiles Takasu, K. & Lewis, W. J. 1996. The role of learning in adult food exploited by an egg parasitoid of pine sawflies. Entomologia Exper- location by the larval parasitoid, Microplitis croceipes (Hymenop- imentalis et Applicata, 115, 217e225. tera: Braconidae). Journal of Insect Behavior, 9, 265e281. Ninkovic, V., Al Abassi, S. & Pettersson, J. 2001. The influence of Tenhumberg, B., Siekmann, G. & Keller, M. A. 2006. Optimal time aphid-induced plant volatiles on ladybird beetle searching behav- allocation in parasitic wasps searching for hosts and food. Oikos, ior. Biological Control, 21, 191e195. 113, 121e131. Olson, D. M., Rains, G. C., Meiners, T., Takasu, K., Tertuliano, M., Turlings, T. C. J. & Benrey, B. 1998. Effects of plant metabolites on Tumlinson, J. H., Wa¨ckers, F. L. & Lewis, W. J. 2003. Parasitic the behavior and development of parasitic wasps. Ecoscience, 5, wasps learn and report diverse chemicals with unique condition- 321e333. able behaviors. Chemical Senses, 28, 545e549. Turlings, T. C. J., Wa¨ckers, F., Vet, L. E. M., Lewis, W. J. & Tumlin- Papaj, D. R. & Lewis, A. C. 1993. Insect Learning: Ecological and son, J. H. 1993. Learning of host-finding cues by hymenopterous Evolutionary Perspectives. New York: Chapman & Hall. parasitoids. In: Insect Learning: Ecological and Evolutionary Perspec- Papaj, D. R. & Vet, L. E. M. 1990. Odor learning and foraging suc- tives (Ed. by D. R. Papaj & A. C. Lewis), pp. 51e78. New York: cess in the parasitoid, Leptopilina heterotoma. Journal of Chemical Chapman & Hall. Ecology, 16, 3137e3150. Turlings, T. C. J., Lengwiler, U. B., Bernasconi, M. L. & Wechsler, Pschorn-Walcher, H. 1982. Symphyta. In: Forstscha¨dlinge Europas. D. 1998. Timing of induced volatile emissions in maize seedlings. Vol. 4 (Ed. by W. Schwenke), pp. 66e103. Hamburg: Parey. Planta, 207, 146e152. Pschorn-Walcher, H. & Eichhorn, O. 1973. Studies on the biology Turlings, T. C. J., Gouinguene´, S., Degen, T. & Fritzsche-Hoballah, and ecology of the egg parasites (Hym. Chalcidoidea) of the pine M. E. 2002. The chemical ecology of plantecaterpillareparasitoid sawfly Neodiprion sertifer (Geoff.) (Hym. Diprionidae) in Central interactions. In: Multitrophic Level Interactions (Ed. by T. Europe. Zeitschrift fu¨r Angewandte Entomologie, 74, 286e318. Tscharntke & B. Hawkins), pp. 148e173. Cambridge: Cambridge Quicke, D. L. J. 1997. Parasitic Wasps. New York: Chapman & Hall. University Press. Said, I., Renou, M., Morin, J. P., Ferreira, J. M. S. & Rochat, D. Vet, L. E. M. & Dicke, M. 1992. Ecology of infochemical use by nat- 2005. Interactions between acetoin, a plant volatile, and phero- ural enemies in a tritrophic context. Annual Review of Entomology, mone in Rhynchophorus palmarum: behavioral and olfactory 37, 141e172. neuron responses. Journal of Chemical Ecology, 31, 1789e1805. Vet, L. E. M. & Groenewold, A. W. 1990. Semiochemicals and learn- Siekmann, G., Tenhumberg, B. & Keller, M. A. 2001. Feeding and ing in parasitoids. Journal of Chemical Ecology, 16, 3119e3135. survival in parasitic wasps: sugar concentration and timing matter. Vet, L. E. M., Van Lenteren, J. C., Heymans, M. & Meelis, E. 1983. Oikos, 95, 425e430. An airflow olfactometer for measuring olfactory responses of Siekmann, G., Keller, M. A. & Tenhumberg, B. 2004. The sweet hymenopterous parasitoids and other small insects. Physiological tooth of adult parasitoid Cotesia rubecula: ignoring hosts for Entomology, 8,97e106. nectar? Journal of Insect Behavior, 17, 459e476. Vet, L. E. M., Lewis, W. J. & Carde´,R.T.1995. Parasitoid foraging Smid, H. M. 2006. Variation in learning of herbivory induced plant and learning. In: Chemical Ecology of Insects (Ed. by R. T. odours by parasitic wasps: from brain to behaviour. In: Chemical Carde´ & W. J. Bell), pp. 65e101. New York: Chapman & Hall. Ecology: from Gene to Ecosystem (Ed. by M. Dicke & W. Takken), Veteli, T. O., Koricheva, J., Niemala, P. & Kellomaki, S. 2006. pp. 89e103. Berlin: Springer. Effects of forest management on the abundance of insect pests e Smith, B. H. 1993. Merging mechanism and adaptation: an on Scots pine. Forest Ecology and Management, 231, 214 ethological approach to learning and generalization. In: Insect 217. Learning: Ecological and Evolutionary Perspectives (Ed. by D. R. Vinson, S. B. 1998. The general host selection behavior of parasit- Papaj & A. C. Lewis), pp. 126e157. New York: Chapman & Hall. oid hymenoptera and a comparison of initial strategies utilized Smith, B. H. 1998. Analysis of interaction in binary odorant by larvaphagous and oophagous species. Biological Control, 11, mixtures. Physiology & Behavior, 65, 397e407. 79e96. Steidle, J. L. M. 1998. Learning pays off: influence of experience on Wa¨ckers, F. L., Bonifay, C. & Lewis, W. J. 2002. Conditioning of host finding and parasitism in Lariophagus distinguendus. Ecological appetitive behavior in the Hymenopteran parasitoid Microplitis Entomology, 23, 451e456. croceipes. Entomologia Experimentalis et Applicata, 103, 135e138. Steidle, J. L. M. & van Loon, J. J. A. 2003. Dietary specialization and Wa¨ckers, F. L., van Rijn, P. C. J. & Bruin, J. 2005. Plant-Provided infochemical use in carnivorous arthropods: testing a concept. Food for Carnivorous Insects: a Protective Mutualism and Its Applica- Entomologia Experimentalis et Applicata, 108, 133e148. tions. Cambridge: Cambridge University Press.