The light brown apple moth, Epiphyas postvittana is a key pest of wine grapes in Australia. Two parasitoids, Dolichogenidea tasmanica and Therophilus unimaculatus, attack the larval stage of this pest. D. tasmanica is dominant in vineyards, whereas T. unimaculatus is mainly active in native vegetation. We sought to understand why they differ in their use of habitats. Plants are a major component of habitats of parasitoids, and herbivore-infested plants influence parasitoid foraging efficiency by their architecture and emission of volatile chemicals. We investigated how different plant species infested by E. postvittana could affect the foraging success of the two parasitoid species in both laboratory and field experiments. Four common host-plant species were selected for this study. In paired-choice experiments to determine the innate foraging preferences for plants, both parasitoid species showed differences in innate search preferences among plant species. The plant preference of D. tasmanica was altered by oviposition experience with hosts that were feeding on other plant species. In a behavioral assay, the two parasitoid species allocated their times engaged in various types of behavior differently when foraging on different plant species. For both parasitoids, parasitism on Hardenbergia violacea was the highest of the four plant species. Significantly more larvae dropped from Myoporum insulare when attacked than from the other three host-plant species, which indicates that parasitism is also affected by interactions between plants and host insects. In vineyards, parasitism by D. tasmanica was significantly lower on M. insulare than on the other three host-plant species, but the parasitism rates were similar among the other three plant species. Our results indicate that plants play a role in the habitat preferences of these two parasitoid species by influencing their foraging behavior, and are likely to contribute to their distributions among habitats.
Citation: Feng Y, Wratten S, Sandhu H, Keller M (2015) Host Plants Affect the Foraging Success of Two Parasitoids that Attack Light Brown Apple Moth Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae). PLoS ONE 10(4): e0124773. https://doi.org/10.1371/journal.pone.0124773
Academic Editor: Nicholas J. Mills, University of California, Berkeley, UNITED STATES
Received: September 10, 2014; Accepted: March 6, 2015; Published: April 20, 2015
Copyright: © 2015 Feng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: The following sources of funding have supported the work: Chinese Scholarship Council (CSC) scholarship Grant Number: 2010630004 and Grape and Wine Research and Development Corporation (GWRDC) PhD Research Scholarship, Grant Number: PH1007 to Yi Feng. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Successful parasitism by parasitoids begins with a series of host-searching behaviors that lead females to locate their potential hosts. This process includes habitat location, host location, and host acceptance . Parasitoid searching behavior is under strong natural selection pressure, because successful foraging is directly linked to reproduction [2,3]. Host-searching behavior determines the efficiency of parasitoids, and thus understanding it is a key element in evaluating their suitability as biological control agents .
Herbivore-infested plants influence the foraging efficiency of parasitoids in various ways [5,6]. For example, herbivore-induced volatiles emitted by the host plant are key signals that lead parasitoids to their hosts [7,8]. Herbivore-infested host plants can selectively attract natural enemies  and, in some cases when different plant species are infested with the same herbivore, some species may attract more parasitoids than the others, resulting in higher parasitism rates under natural conditions . In addition to plant volatiles, plant architecture can also influence the interactions between the parasitoids and their hosts [11–16]. Plant morphological characteristics such as plant-surface structural complexity , presence of dense trichomes [18,19], and leaf surface area [14,20,21] can significantly influence the success rates of parasitoids or predators in finding their hosts. To understand how plants affect parasitoid foraging efficiency, the effects of plant attributes including both plant volatiles and other characteristics like architecture should be considered.
Parasitoids may also rely on both innate mechanisms and learned cues associated with host availability during foraging . Through learning, parasitoids can adaptively optimize their foraging efficiency by altering their innate preferences . The effects of learning on the ability of parasitoids to locate hosts have been documented in both laboratory [24,25] and field studies . For instance, foraging experience on different herbivore-infested cabbage varieties can lead generalist parasitoids to have a preference for the herbivore-infested plant species that they have experienced . However, when indigenous parasitoids forage for hosts on a wide range of plant species involving both native and exotic plants, it is not known whether previous oviposition experience on one plant species will influence their subsequent foraging preference in favor of the same species.
In agro-ecosystems, native plant species happen to grow within or adjacent to the crop plants, which are mostly non-native introduced species. To investigate the effect of plant species on the foraging efficiency of generalist parasitoids, both the plant attributes and the adaptive learning behavior of the parasitoids should be considered.
In this study, we investigated the foraging behavior of Dolichogenidea tasmanica (Cameron) (Hymenoptera: Braconidae) and Therophilus unimaculatus (Turner) (Hymenoptera: Braconidae) (Fig 1). Both species are indigenous to Australia and are solitary, koinobiont, generalist endoparasitoids [27–29]. They both attack the light brown apple moth, Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae), which is a native Australian, leaf rolling, polyphagous, multivoltine moth. Epiphyas postvittana is the key insect pest of grapevines in Australia , and it also attacks plants from 123 genera in 55 plant families in this country. Among these plant species, 22 are from native genera, while 101 are from exotic genera . Approximately 25 hymenopteran parasitoids are reported to be associated with E. postvittana in Australia , however, D. tasmanica and T. unimaculatus are the predominant parasitoids of E. postvittana. The former is the most abundant larval parasitoid in vineyards, while the latter one is most common in the adjacent vegetation . It is not known why one parasitoid species is more active in vineyards while the other one is not. Plants could play a key role among a number of factors that affect the activity of these parasitoids in vineyard ecosystems. However, it is not known how different plant species in and around the vineyards affect their foraging behavior and habitat associations of these two parasitoids.
These actively foraging parasitoids are generalists that attack a range of leafrollers (Lepidoptera: Tortricidae). Photos by Michael Keller (a), and Yi Feng (b).
We investigated (1) whether female D. tasmanica and T. unimaculatus have innate search preference for different host plant species infested with larval E. postvittana; (2) whether previous oviposition experience alters the host-plant preferences of the parasitoids; (3) how different plant species infested with E. postvittana affect the behavior and foraging efficiency of the two parasitoids; and (4) whether parasitism is affected by different plant species in vineyards. We first tested the in-flight preference of both parasitoid species for host-infested plants with dual-choice wind tunnel assays. Four representative plants species, which included both native and exotic species, were selected. We then tested whether previous oviposition experience on one plant species could influence the in-flight preference of D. tasmanica for host plants. Next, we investigated the searching behavior of two parasitoids on four plant species. Finally, a field experiment was conducted to determine whether parasitism of larval E. postvittana was influenced by their host plants in vineyards. Overall, we gained insights into how plant species affect the foraging success of two generalist parasitoids that attack E. postvittana.
Materials and Methods
Ethics Statement: Permission to conduct the field experiment was granted by the vineyard managers Geoff Hardy and David Hamilton. No permit was required for the laboratory studies.
Insects and Plants
An artificial diet was used to rear E. postvittana  at 22 ± 2°C under a 12 L:12 D light: dark cycle in an insect-rearing room. This culture has been maintained for 200 generations, with the annual addition of field-collected individuals to maintain genetic diversity. The colonies of D. tasmanica and T. unimaculatus were originally established from parasitized leafrollers collected in a vineyard (35°16'05'' S, 138°37'10'' E) near Adelaide, Australia in November 2011. These parasitoids were reared on larval E. postvittana that fed on plantain (Plantago lanceolata L.) and were maintained in cages at 23 ± 2°C with a relative humidity of 60 ± 10% under a 14 L: 10 D light: dark cycle. Naturally occurring larval E. postvittana were collected from the field (35°58'18'' S, 138°38'32'' E) every two months and the newly emerged adult parasitoids were added to the respective colonies.
We chose four plant species that have been reported to be common host-plants for leafrollers in Australia  representing three categories: (1) an introduced economic crop, wine grape, Vitis vinifera L., cv. Chardonnay, which is highly susceptible to attack by E. postvittana ; (2) two Australian native perennial plants, Hardenbergia violacea (Schneev.) Stearn and Myoporum insulare R. Br.; and (3) an exotic ground cover species, plantain, Plantago lanceolata L. These plant species differ in their architecture and the level of protection available to larval E. postvittana.
Experimental plants were grown in containers. P. lanceolata was grown from seed three months prior to the experiment. The Chardonnay grape vines were grown from pencil-sized cuttings collected from a vineyard during winter. Native plants (H. violacea and M. insulare) (~20 cm high) were purchased from a Nursery. For the laboratory experiments, all plants were grown individually in UC soil mix  in plastic pots (50 mm × 50 mm × 120 mm). For the field experiment, all plants were grown individually in UC soil mix and cocopeat potting mix at a ratio of 1:1 in nursery bags (300 mm × 120 mm × 150 mm) in a glasshouse. Plants were placed in a field cage two weeks before the onset of field experiments to allow them to acclimatize to natural conditions.
We used two- or three-day old mated female parasitoids. Newly formed parasitoid cocoons were collected and held individually in 100 ml plastic cups, each with a drop of honey and a water-soaked cotton dental wick. The newly emerged female parasitoids were caged with five males for 24 hours, with a drop of honey and water-soaked cotton dental wick, to ensure mating. The mated females were then isolated in glass vials (18 mm diam × 50 mm) with a drop of honey. Immediately before release, the individual parasitoids were primed by exposing them to feces collected from E. postvittana reared on an artificial diet in a Petri dish (80 mm diam). This stimulated the parasitoids with host-related cues that were not from any of the experimental plants. Individual parasitoids were then transferred to a clean vial (same as above) for release in the wind tunnel. The bottom half of the vial was filled with cotton to ensure the parasitoid did not move to the bottom and stay there.
An experiment was conducted to test the initial in-flight preference of the parasitoids for volatiles from different plants that were damaged by E. postvittana. There were four plant species, and therefore six pairs of volatile sources that were tested for each parasitoid species. Both parasitoid species were tested in dual-choice situations in which two volatile sources were placed in pairs in a wind tunnel at a wind speed of 20 cm/s at 21 ± 2°C (Fig 2; see  for wind tunnel details). To strengthen the volatile emissions and ensure continuous host-feeding damage on the plant leaves during the experiment, leaves of two plant species were infested with 20 second-instar E. postvittana. To reduce variations in morphology, texture, color, and structure and to ensure each parasitoid had a free and equal choice to fly to either of the target plants, the leaves of both plant species were placed in a metal screen tea ball (5.5 cm diam). Pilot tests indicated that only D. tasmanica would fly to host infested leaves that were enclosed in a tea ball. Therefore, to test the inflight preference of T. unimaculatus, plants infested with 20 second instar E. postvittana 24 h before the test were used. While different methods were used to test the innate preferences for host infested plants by these parasitoid species, this difference did not compromise our overall aim, which was to determine how host infested plants might differentially attract these parasitoids. Care was taken that all volatile sources from the same plant pair were used on the same day. To avoid plant-position bias effects, the positions of the two tea balls or the intact plants were randomized and with equal numbers of each test plant in each position. A single parasitoid was released from a glass vial on a stand 25 cm downwind from the tea balls or plants at approximately the same height, which were 25 cm above the floor of the wind tunnel and separated by 5 cm. The observation for each parasitoid lasted a maximum of five minutes. A ‘choice’ was recorded when a female landed on a tea ball or an intact plant. Wasps that did not respond within five minutes or landed elsewhere were recorded as ‘no response’. The experiments were conducted between 9:00 am and 4:00 pm, and each parasitoid was tested only once. Thirty-six parasitoids of each species were tested for each plant pair.
Effects of Learning on Host-Plant Preference
An experiment was conducted with D. tasmanica to examine whether previous oviposition experience on host-infested plants alters the subsequent preference for host-induced plant volatiles. To provide female parasitoids with multiple oviposition experiences, they were allowed five sequential ovipositions on leaves of grape or P. lanceolata. A choice experiment was then conducted with an experimental design that was similar to the previous choice test. The parasitoids were observed in the wind tunnel to determine their landing preference between two plant species, of which one was the species of their previous oviposition experience. The experiment was conducted with 36 experienced parasitoids for each pair of plants and 36 naive parasitoids tested with the same pair of plants as a control, which also served to validate the results of the previous experiment. Due to insufficient numbers of T. unimaculatus, this parasitoid species was not included in this part of our study.
To determine if parasitoid behavior varies among host plants, we also observed their foraging activities on the four plant species, which was influenced by the combined effects of all plant characteristics (plant chemistry and structure). This study was conducted in a wind tunnel (Fig 3). An individual plant infested with two second-instar E. postvittana was placed upwind of the parasitoid in the wind tunnel. To avoid the parasitoid spending excess time on the test plant, a second host plant of the same species and condition as the test plant was placed 40 cm upwind during each observation to provide an alternative landing place for the parasitoid. The second plant was also infested with two second-instar E. postvittana. Before the experiment, all of the leaves on the test plants were examined to check for the position and number of host larvae, host damage, and frass on the plants. A single parasitoid was released from a glass vial 25 cm downwind from the host infested plant. The foraging behavior of individual parasitoids on the downwind host-infested plant was then observed, with the observations lasting until the parasitoid left the plant to either another location in the wind tunnel or the alternative plant. For both parasitoid species, twenty wasps were observed for each host-plant species, and each parasitoid was observed only once. The order of testing the plant species was randomized.
An individual plant infested with two second-instar larval E. postvittana was placed upwind of the parasitoid in the wind tunnel. To avoid the parasitoid spending excess time on the test plant, a second host plant of the same species and condition as the test plant was placed further upwind during each observation to provide an alternative landing place for the parasitoid.
Based on preliminary observations, a catalogue of female behavior for D. tasmanica and T. unimaculatus searching was constructed (Table 1). The mean duration for each type of behavior was calculated from when a parasitoid first landed on the host plant. According to a pilot experiment, a common host defensive behavior to avoid parasitoid attack is to drop from the plant. Therefore, the dropping behavior of the host larvae was also recorded, as were the number of larvae that were stung. Each day, observations were conducted between 10:00 am and 4:00 pm. Parasitoid behavior was recorded with the Observer XT ver. 11 software package (Noldus Information Technology B.V., Wageningen, the Netherlands). The egg hatching time of D. tasmanica and T. unimaculatus are about two and four days after egg laying, respectively . Therefore, the larvae stung by the D. tasmanica and T. unimaculatus were dissected two and four days after the experiment, respectively, to determine the parasitism rate.
Field Experiment—Parasitism of E. postvittana on Four Plant Species
To evaluate whether the four plant species influence the levels of parasitism by D. tasmanica and T. unimaculatus in vineyards, a field experiment was conducted at two sites (vineyard A: 35°13' S, 138°39' E; vineyard B: 35°16' S, 138°37' E) and repeated twice at each vineyard between February and March 2013. An orthogonal split-plot design was used where the vineyards were considered random blocks, the plant species were the main-plot factor, and the repeated visits were the split-plot factor. In each vineyard, the four different species of potted plants infested with first-instar E. postvittana (around 20 larval E. postvittana on each plant) were placed in five sets of quadruplets (sample unit = 3 plants/species) inside the vineyard and 20 m from the border (Fig 4). At each site, 60 plants were placed in the field and left for two weeks of free access to wild parasitoids. The plants were placed in the two vineyards on consecutive days. After two weeks, the plants were removed and replaced with a fresh pair of plants. The larvae on each plant were collected and reared in 440 ml plastic containers at 22 ± 2°C under a 14 L:10 D light/dark cycle in an insect-rearing room. The parasitism rate and fate of the larvae (dead, parasitized, or pupated) were recorded.
To determine which host plant was preferred by the parasitoids, the choice within an experiment was analyzed using a binomial test, with 0.5 as the null hypothesis. To examine if learning can alter the landing preference of the parasitoids, chi-squared tests with Cochran's correction for continuity  were used to compare the landing choices of the naive/control and experienced wasps.
The mean proportion of time devoted to each type of behavior after landing on a plant was calculated for each parasitoid species on each plant species. The proportion of time that wasps were engaged in each behavior was calculated for each individual, and differences among the plant species for each parasitoid were analyzed with Kruskal-Wallis tests (IBM SPSS Statistics v. 20, IBM-SPSS Inc., Chicago, IL).
In addition, Chi-square tests were also used to determine if differences in parasitism rates and host defensive behaviors (larvae dropping) among four plant species were statistically significant. If the null hypothesis was not rejected, a retrospective power analysis for Chi-square was carried out (Functions CHISQ_POWER and CHISQ_SIZE, default power = 0.8, www.real-statistics.com) to determine the required sample size with the significance level of 0.05.
To analyse the factors affecting parasitism in the experimentally introduced E. postvittana, the data from field experiments were modelled with orthogonal split-plot general linear models with the GLM procedure in the statistical package GenStat for windows 15th Edition (VSN International, Hemel Hempstead, UK.). For all experiments, the fractions of larvae that were parasitised by all parasitoids on each plant species in each quadruplet were calculated from the pooled numbers; these numbers were treated as the dependent variables. The arcsine transformation  was used to analyse the parasitism data. The Bonferroni adjustment was used for post-hoc multiple comparisons of means. The level of significance was set at 0.05.
Parasitoids had varied preferences for landing in the dual-choice experiments. Female D. tasmanica preferred to land on one species combination over another in four of the six plant-host-complex pairings (Fig 5a). H. violacea was preferred over M. insulare, grape and P. lanceolata, and M. insulare was preferred over grape. T. unimaculatus females showed preference in three of the six pairings (Fig 5b). H. violacea was preferred over grape and P. lanceolata, and M. insulare was preferred over grape.
Within each choice test, the per cent of parasitoids that made no choice is shown at the right. Host plants: P, P. lanceolata; G, grape; H, H. violacea; M, M. insulare. Asterisks indicate a significant difference between the targets (binomial test; *P < 0.05, **P < 0.005). NS, not significant.
Effects of Learning on Host Plant Preference
Oviposition experience on a plant affected subsequent plant preferences in D. tasmanica. In the experiments that involved experience on P. lanceolata, the change in preference between control and experienced wasps was statistically significant only when the comparison between them involved the choice between P. lanceolata and grape (Fig 6a). However, there was a statistical change in the degree of preference between species in all paired groups of control and experienced wasps. For example, naive wasps displayed no preference between M. insulare and P. lanceolata, but wasps that had experience on P. lanceolata significantly preferred to land on it. Likewise, naive wasps that were presented with choice between H. violacea and P. lanceolata preferred to land on H. violacea, but this preference was not displayed by wasps that had experience on P. lanceolata. Oviposition experience on grapes led to more frequent landing on grapes by D. tasmanica in the presence of each of the other three plant species (Fig 6b). Interestingly, the preference of naive D. tasmanica for H. violacea over grape was reversed after females had oviposition experiences on hosts that feeding on grapes.
The female D. tasmanica had previous oviposition experience on Plantago lanceolata (a) or grape (b). Host plants: P, P. lanceolata; G, Grape; H, H. violacea; M, M. insulare. Control groups were naive parasitoids. Within each choice test, the per cent of parasitoids that made no choice is shown at the right. Asterisks on the left side of the table indicate a significant difference between the targets (binomial test; *P < 0.05, **P < 0.005). Asterisks on the right side of the table indicate a significant difference between the experimental and control groups (χ2 tests with Cochran’s correction for continuity; *P < 0.05). NS, not significant.
The behavioral profiles of both parasitoids varied among the four host-plant species (Fig 7; raw data in S1 Table). The fractions of time D. tasmanica engaged in stinging, probing and grooming differed among plant species, while the fractions of time T. unimaculatus allocated to pulling, antennating and grooming differed among species. The parasitism rate of T. unimaculatus on H. violacea was higher than on the other three plant species. The parasitism rate by D. tasmanica on H. violacea was higher than on other plant species but this difference was not statistical significant, possibly due to low statistical power (Fig 8a). The most common way that larval E. postvittana avoided attack was to drop from the plant on a silk thread when contacted by a parasitoid. From our laboratory observations, more larvae dropped from M. insulare than from the other plants (Fig 8b). Among the dropping hosts pooled among all plants, significantly more were parasitized by D. tasmanica than by T. unimaculatus (Fig 9). The two parasitoid species responded differently toward the escaping hosts. When encountering an escaping host hanging from a silk strand, D. tasmanica immediately followed the larva or attacked it on the thread, or searched for few seconds, flew to locate it, and then attacked the larvae on the thread. If the host dropped to the ground, some of the D. tasmanica flew or walked to the ground and searched for the host. In contrast, T. unimaculatus displayed hauling behavior when encountering a dropping larva, continuously reeling the silk thread with its legs up to pull the hanging larva back to the plant, even though in many cases the larva would drop to the ground. No T. unimaculatus was observed following the dropping host or searching for the larvae on the ground. D. tasmanica found and attacked the dropping hosts more successfully than did T. unimaculatus (Fig 9).
Behaviors (see Table 1 for definitions): An, antennating; Fl, flying; Gr, grooming; Pr, probing; Pu, pulling; Sta, stationary; Sti, stinging; Wa, walking. Host plants: P, P. lanceolata; G, Grape; H, H. violacea; M, M. insulare. Bold behavior characters indicate significant differences among plant species (Kruskal Wallis tests; P < 0.05).
The number of larvae that dropped/total differed significantly among four host plant species (D. tasmanica, χ2 = 10.80, P = 0.013; T. unimaculatus, χ2 = 18.15, P = 0.0004). Number of larvae parasitised was not different among host plant species for D. tasmanica (χ2 = 5.27, P = 0.153). Chi-square power analysis (default power = 0.8) indicated that the sample size must be increased by 2.06 times before the experimental results would be likely to produce statistically significant differences. No. of larvae parasitised was higher on H. violacea than the other three host plant species for T. unimaculatus (χ2 = 8.51, P = 0.037). Numbers within each bar indicate the sample sizes.
Field Experiment—Parasitism of E. postvittana on Four Plant Species
The number of larval E. postvittana that were found on each three-pot group of plant species differed significantly among the four plant species (Fig 10a; F3,76 = 70.06, P < 0.05; raw data in S2 Table). A low number of larval E. postvittana was recovered from M. insulare compared with other plant species in both replications.
The larvae were feeding on grape, H. violacea, P. lanceolata, and M. insulare. Data are expressed as overall means ± standard error.
Only two parasitoid species, D. tasmanica and T. unimaculatus, were recovered from the larval E. postvittana on plants placed in the vineyards. No T. unimaculatus were collected from larvae on M. insulare. The angular-transformed parasitism rates for D. tasmanica and overall parasitism differed among plant species (Fig 10b; D. tasmanica, F3, 76 = 12.41, P < 0.05; overall parasitism, F3, 76 = 34.64, P < 0.05). However, the angular-transformed parasitism rates for T. unimaculatus were not different among plant species (Fig 10b; T. unimaculatus, F3, 76 = 2.52, P = 0.234), which reflects the low and inconsistent appearance of this species between sites and dates. Overall parasitism on M. insulare was significantly lower than on grape and H. violacea, but there were no detectable differences in parasitism by D. tasmanica among plant species when Bonferroni-test adjustments were made to multiple comparisons.
The results of this study indicate that herbivore infested plants affect the foraging behavior and efficiency of both D. tasmanica and T. unimaculatus. Although we did not directly measure and compare the herbivore-induced plant volatile profiles of the plant species we tested, the effect of host induced plant volatiles were directly observed through comparing the responses of the parasitoids to paired host-infested plants. The results indicate both parasitoids have innate preferences for plant species. The Australian native plant, H. violacea, was the most attractive species to both parasitoids. Studies have demonstrated that among a range of plant species infested with the same herbivore, the host-induced volatiles differ in both specificity and quantity, and affect the natural enemies differently [9,40,41]. The native plant H. violacea was more attractive to the parasitoids, but the exotic plants P. lanceolata and grape, which have established recent interactions with the second and third trophic levels, also attracted the parasitoids. This suggests adaptability by these generalist parasitoids as well as innate responses to common characteristics of plants. In this study, the plant species on which parasitoids were reared could have potentially affected their behavior . However, our main aim was to evaluate how plant species could generally affect the behavior of parasitoids that are foraging in a diverse landscape. We would not expect any potential bias from rearing on a selected plant species to affect this broader evaluation.
Previous oviposition experience with hosts that were feeding on grape and P. lanceolata both affected the subsequent plant preference of D. tasmanica in the dual-choice tests (Fig 6a–6b). When all statistical analyses are considered together, the results indicate that D. tasmanica is more likely to fly to a particular plant species after it has experience with parasitizing a host on it. The magnitude of such shifts in preference is likely to depend on the degree of innate preference, the level of experience and the characteristics of the plants involved. From these results, we hypothesize that T. unimaculatus can also learn to associate hosts with some plants as a result of experience. Assuming this is the case, the experiences of both species could reinforce their preferences with hosts on innately preferred and common host plant species in or near vineyards. On the other hand, experience on previous less preferred plants may lead to a reduction in preference for a host plant that is preferred by a naive parasitoid.
D. tasmanica and T. unimaculatus allocated their times engaged in various types of behavior differently when foraging on different plant species. In addition, D. tasmanica spends relatively more time antennating and probing and less time stinging than T. unimaculatus (Fig 7). These results indicate that a combination of plant characteristics influences the foraging behavior of these parasitoids. Many studies have demonstrated that plant structure can affect parasitoid foraging efficiency . The size, heterogeneity and connectivity of plants have been shown to affect parasitoid foraging success, as confirmed both by modelling the impact of plant structure on parasitism rates based on artificial plants and by experiments with real plants . The probability of parasitoids and predators encountering a host or prey generally decreases with an increase in plant structural complexity [11,14,16,20,45,46], plant size , and plant surface area or volume [48–50].
Our results also indicate that the plant species affect the defensive behavior of the host (Fig 8), and therefore indirectly affect the foraging efficiency of parasitoids (Fig 9). Host larvae falling from plants may be at a higher risk of encountering other predators or parasitoids . Because of this, plant species that facilitate escaping behavior in the field may not support large numbers of parasitoids.
D. tasmanica was found to be the dominant species that parasitizes larval E. postvittana in vineyards. Both D. tasmanica and T. unimaculatus were found in the field experiment, but D. tasmanica parasitised the majority of the sentinel larval E. postvittana in the vineyards. Parasitism by D. tasmanica was consistent between sites and dates, while parasitism by T. unimaculatus was inconsistent and low (Fig 10b). This is in line with a previous study that indicated the dominance of D. tasmanica in vineyards . Under field conditions, host-related chemical cues are the main long-distance attractor for parasitoids, which could influence their foraging efficiency. Plants could be a key factor that affects the activity of parasitoids in certain habitats. Studies have suggested that plants which are attractive to parasitoids are associated with a higher parasitism rate under field conditions . There is independent evidence that parasitism by D. tasmanica varies significantly among plant species , which is consistent with the results of our field experiment. Moreover, our behavioral assays indicated that plants vary in their level of attraction to females, and this can be influenced by a wasp's experience. In one behavioral assay, parasitism rates did not differ among plant species, but this non-significant result is likely to be due to low statistical power (Fig 8a). This indicates that differences among plants may not be pronounced when some species combinations are compared.
Research has indicated that successful foraging on certain host plants can narrow a parasitoid’s foraging range through learning . The field experiment was carried out in vineyards, where the main background plants were grapes and P. lanceolata was a common ground cover plant. Although H. violacea could be the preferred host plant species of the naive wasp, the parasitoids may gain experiences more frequently after attacked hosts feeding on the more abundant and naturally occurring grape or P. lanceolata in vineyards and thus strengthen their habitat preference. This could happen for both D. tasmanica and T. unimaculatus that were attracted to vineyards. However, when both parasitoids are active in a vineyard, their competitive interactions would influence their foraging success and abundance . In this case, the more abundant parasitoid species should be found to have a competitive advantage over its competitors.
Results from our field experiment showed the number of larval E. postvittana recovered from the four different plant species varies. It is noticeable that much smaller numbers of larval hosts were recovered from M. insulare compared to the other three plant species (Fig 10a). Studies have indicated that plants infested with a relatively high density of host larvae should attract more parasitoids . Therefore, the smaller number of host larvae could have negatively affected the parasitism of E. postvittana on M. insulare. It is also possible that larvae could not defend themselves as effectively from parasitoids and predators on M. insulare. In the case of parasitoid attack, larvae that dropped from the plant may have failed to return to it more frequently than on other host plant species.
In conclusion, plants influence host availability and attract the two parasitoids differently. Different plant species provide different levels of protection for larval E. postvittana and thereby affect the foraging behavior and efficiency of parasitoids that attack it. Results of our field experiment indicate D. tasmanica is the dominant parasitoid in vineyards, and plant species affect parasitism by this parasitoid species. Putting these effects together, we conclude plants are likely to affect the habitat preferences and distributions of these parasitoids that share the same hosts. In this study we investigated tri-trophic interactions involving two Australian native plants, H. violacea and M. insulare, as well as the introduced exotics, P. lanceolata and grape. This approach revealed how plants generally affect the parasitoid-host interactions in real vineyard ecosystems, in which numerous plant species are involved in the interactions between the herbivore and its associated parasitoid species.
Future research should address the question of whether adding specific plant species that are preferred by parasitoids could increase the suppression of E. postvittana in vineyards. To strengthen the activity of parasitoids that are already active in vineyards or attract more parasitoids into vineyards, it is necessary to evaluate a wider range of supplementary plant species, especially native and perennial species. Understanding the interactions at the tri-trophic level of plant, pest and natural enemies should help us “ecologically engineer” vineyards  to promote the activity of these parasitoids and therefore enhance the ecosystem service of biological control .
We thank Maryam Yazdani for assistance with rearing the insects. Comments from Nicholas J Mills and two anonymous reviewers improved the manuscript considerably.
Conceived and designed the experiments: YF MK SW HS. Performed the experiments: YF. Analyzed the data: YF MK. Contributed reagents/materials/analysis tools: YF MK. Wrote the paper: YF MK SW HS.
- 1. Vinson SB. Host selection by insect parasitoids. Annu Rev Entomol. 1976; 21: 109–133.
- 2. Vet LEM, van Alphen JJM. A comparative functional approach to the host detection behaviour of parasitic wasps. 1. A qualitative study on Eucoilidae and Alysiinae. Oikos. 1985; 44: 478–486.
- 3. Godfray HCJ. Parasitoids: behavioral and evolutionary ecology. New Jersey: Princeton University Press; 1994.
- 4. Luck RF. Evaluation of natural enemies for biological control: a behavioral approach. Trends Ecol Evol. 1990; 5: 196–199. pmid:21232353
- 5. Price PW, Bouton CE, Gross P, McPheron BA, Thompson JN, Weis AE. Interactions among three trophic levels: influence of plants on interactions between insect herbivores and natural enemies. Annu Rev Ecol Syst. 1980; 11: 41–65.
- 6. Lill J, Marquis R, Ricklefs R. Host plants influence parasitism of forest caterpillars. Nature. 2002; 417: 170–173. pmid:12000959
- 7. Vet LEM, Dicke M. Ecology of infochemical use by natural enemies in a tritrophic context. Annu Rev Entomol. 1992; 37: 141–172.
- 8. Geervliet JBF, Vet LEM, Dicke M. Volatiles from damaged plants as major cues in long-range host-searching by the specialist parasitoid Cotesia Rubecula. Entomol Exp Appl. 1994; 73: 289–297.
- 9. De Moraes C, Lewis W, Pare P, Alborn H, Tumlinson J. Herbivore-infested plants selectively attract parasitoids. Nature. 1998; 393: 570–573.
- 10. Poelman EH, Oduor AMO, Broekgaarden C, Hordijk CA, Jansen JJ, Van Loon JJA, et al. Field parasitism rates of caterpillars on Brassica oleracea plants are reliably predicted by differential attraction of Cotesia parasitoids. Funct Ecol. 2009; 23: 951–962.
- 11. Geitzenauer HL, Bernays EA. Plant effects on prey choice by a vespid wasp, Polistes arizonensis. Ecol Entomol. 1996; 21: 227–234.
- 12. Bergman JM, Tingey WM. Aspects of interaction between plant genotypes and biological control. Bulletin of the ESA. 1979; 25: 275–279.
- 13. Cortesero A, Stapel J, Lewis W. Understanding and manipulating plant attributes to enhance biological control. Biol Control. 2000; 17: 35–49.
- 14. Carter M, Sutherland D, Dixon A. Plant structure and the searching efficiency of coccinellid larvae. Oecologia. 1984; 63: 394–397.
- 15. Letourneau D. Mechanisms of predator accumulation in a mixed crop system. Ecol Entomol. 1990; 15: 63–69.
- 16. Grevstad FS, Klepetka BW. The influence of plant architecture on the foraging efficiencies of a suite of ladybird beetles feeding on aphids. Oecologia. 1992; 92: 399–404.
- 17. Andow D, Prokrym D. Plant structural complexity and host-finding by a parasitoid. Oecologia. 1990; 82: 162–165.
- 18. Belcher DW, Thurston R. Inhibition of movement of larvae of the convergent lady beetle by leaf trichomes of tobacco. Environ Entomol. 1982; 11: 91–94.
- 19. Kauffman WC, Kennedy GG. Relationship between trichome density in tomato and parasitism of Heliothis spp. (Lepidoptera: Noctuidae) eggs by Trichogramma spp. (Hymenoptera: Trichogrammatidae). Environ Entomol. 1989; 18: 698–704.
- 20. Kareiva P, Sahakian R. Tritrophic effects of a simple architectural mutation in pea plants. Nature. 1990; 345: 433–434.
- 21. Keller MA. Influence of leaf surfaces on movements by the hymenopterous parasitoid Trichogramma exiguum. Entomol Exp Appl. 1987; 43: 55–59.
- 22. Vet LEM, Lewis WJ, Cardé RT. Parasitoid foraging and learning. In: Cardé RT, Bell WJ, editors. Chemical Ecology of Insects 2. Springer USA. 1995.
- 23. Wackers FL, Lewis WJ. Olfactory and visual learning and their combined influence on host site location by the parasitoid Microplitis croceipes (Cresson). Biol Control. 1994; 4: 105–112.
- 24. Geervliet JBF, Vreugdenhil AI, Dicke M, Vet LEM. Learning to discriminate between infochemicals from different plant-host complexes by the parasitoids Cotesia glomerata and C. rubecula. Entomol Exp Appl. 1998; 86: 241–252.
- 25. Lewis WJ, Takasu K. Use of learned odours by a parasitic wasp in accordance with host and food needs. Nature. 1990; 348: 635–636.
- 26. Papaj DR, Vet LE. Odor learning and foraging success in the parasitoid, Leptopilina heterotoma. J Chem Ecol. 1990; 16: 3137–3150. pmid:24263300
- 27. Dumbleton LJ. Apanteles tasmanica Cam.: a Braconid parasite of leaf-roller larvae. New Zeal J Sci Tec. 1935; 17: 572–576.
- 28. Charles JG, Walker JTS, White V. Leafroller phenology and parasitism in Hawkes bay, New Zealand, canefruit gardens. New Zeal J Crop Hort. 1996; 24: 123–131.
- 29. Stevens NB, Austin AD, Jennings JT. Diversity, distribution and taxonomy of the Australian agathidine genera Camptothlipsis Enderlein, Lytopylus Foerster and Therophilus Wesmael (Hymenoptera: Braconidae: Agathidinae). Zootaxa. 2011; 2887: 1–49.
- 30. Scholefield P, Morison J. Assessment of economic cost of endemic pests & diseases on the Australian grape & wine industry(GWR 08/04). Final report to the Grape & Wine Research & Development Corporation, Adelaide, SA, Australia.2010. Available: http://research.agwa.net.au/completed_projects/assessment-of-economic-cost-of-endemic-pests-and-diseases-on-the-australian-grape-and-wine-industry/. Accessed 14 February 2015.
- 31. Suckling DM, Brockerhoff EG. Invasion biology, ecology, and management of the light brown apple moth (Tortricidae). Annu Rev Entomol. 2010; 55: 285–306. pmid:19728834
- 32. Paull C, Austin AD. The hymenopteran parasitoids of light brown apple moth, Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae) in Australia. Aust J Entomol. 2006; 45: 142–156.
- 33. Paull C. The ecology of key arthropods for the management of Epiphyas postvittana (Walker) (Lepidoptera: Tortricidae) in Coonawarra vineyards, PhD. Thesis, The University of Adelaide. 2007. Available: http://hdl.handle.net/2440/47971. Accessed 14 February 2015.
- 34. Yazdani M, Feng Y, Glatz R, Keller MA. Host stage preference of Dolichogenidea tasmanica (Cameron, 1912)(Hymenoptera: Braconidae), a parasitoid of Epiphyas postvittana (Walker, 1863) (Lepidoptera: Tortricidae). Austral Entomology. 2014.
- 35. Paull C, Schellhorn N, Austin A. Response to host density by the parasitoid Dolichogenidea tasmanica (Hymenoptera: Braconidae) and the influence of grapevine variety. B Entomol Res. 2013; 104: 1–9.
- 36. Matkin O, Chandler P. The UC type soil mixes. In: Baker KF, Chandler PA, editors. UC system for producing healthy container-grown plants. Calif Agric Exp Stn Man. 1957; pp. 332.
- 37. Keller MA. Responses of the parasitoid Cotesia rubecula to its host Pieris rapae in a flight tunnel. Entomol Exp Appl. 1990; 57: 243–249.
- 38. Feng Y Wratten S, Sandhu H, Keller MA. Interspecific competition between two generalist parasitoids that attack Epiphyas postvittana (Lepidoptera: Tortricidae). B Entomol Res. 2015;
- 39. Zar JH. Biostatistical Analysis. 2nd edn. Englewood Cliffs, New Jersey: Prentice-Hall. 1999.
- 40. Geervliet JBF, Vet LEM, Dicke M. Innate responses of the parasitoids Cotesia glomerata and C. rubecula (Hymenoptera: Braconidae) to volatiles from different plant-herbivore complexes. J Insect Behav. 1996; 9: 525–538.
- 41. Geervliet JBF, Posthumus MA, Vet LEM, Dicke M. Comparative analysis of headspace volatiles from different caterpillar-infested or uninfested food plants of Pieris species. J Chem Ecol. 1997; 23: 2935–2954.
- 42. Herard F, Keller M, Lewis W, Tumlinson J. Beneficial arthropod behavior mediated by airborne semiochemicals. IV. Influence of host diet on host-oriented flight chamber responsed of Microplitis demolitor Wikinson. J Chem Ecol. 1988; 14: 1583–1596. pmid:24276431
- 43. Gingras D, Boivin G. Effect of plant structure, host density and foraging duration on host finding by Trichogramma evanescens (Hymenoptera: Trichogrammatidae). Environ Entomol. 2002; 31: 1153–1157.
- 44. Gingras D, Dutilleul P, Boivin G. Modeling the impact of plant structure on host-finding behavior of parasitoids. Oecologia. 2002; 130: 396–402.
- 45. Bond AB. The foraging behaviour of lacewing larvae on vertical rods. Anim Behav. 1983; 31: 990–1004.
- 46. Mackauer M, Völkl W. Regulation of aphid populations by aphidiid wasps: does parasitoid foraging behaviour or hyperparasitism limit impact? Oecologia. 1993; 94: 339–350.
- 47. Ables J, Morrison R. Effect of cotton plant size, host egg location, and location of parasite release on parasitism by Trichogramma pretiosum. Southwest Entomol. 1980; 5: 261–264.
- 48. Bigler F, Brunetti R. Biological control of Ostrinia nubilalis Hbn. by Trichogramma maidis Pint. et Voeg. on corn for seed production in southern Switzerland. J Appl Entomol. 1986; 102: 303–308.
- 49. Burbutis PP, Koepke CH. European corn borer control in peppers by Trichogramma nubilale. J Econ Entomol. 1981; 74: 246–247.
- 50. Kanour WW, Burbutis PP. Trichogramma nubilale (Hymenoptera: Trichogrammatidae) field releases in corn and a hypothetical model for control of European corn borer (Lepidoptera: Pyralidae). J Econ Entomol. 1984; 77: 103–107.
- 51. Paull C, Schellhorn NA, HilleRisLambers R, Austin AD. Escape from parasitoids leave larvae vulnerable to predators and has unexpected outcomes for pest suppression. Basic Appl Ecol. 2012; 13: 542–550.
- 52. Suckling DM, Burnip GM, Walker JTS, Shaw PW, McLaren GF, Howard CR, et al. Abundance of leaf rollers and their parasitoids on selected host plants in New Zealand. New Zeal J Crop Hort. 1998; 26: 193–203.
- 53. Ishii Y, Shimada M. The effect of learning and search images on predator–prey interactions. Popul Ecol. 2010; 52: 27–35. pmid:20416095
- 54. Girling RD, Stewart-Jones A, Dherbecourt J, Staley JT, Wright DJ, Poppy GM. Parasitoids select plants more heavily infested with their caterpillar hosts: a new approach to aid interpretation of plant headspace volatiles. P Roy Soc B-Biol Sci. 2011; 278: 2646–2653. pmid:21270031
- 55. Gurr G, Wratten SD, Altieri MA. Ecological engineering for pest management: advances in habitat manipulation for arthropods. CSIRO Publishing. 2004.
- 56. Wratten S, Sandhu H, Cullen R, Costanza R. Ecosystem services in agricultural and urban landscapes. Wiley-Blackwell, Oxford, UK. 2013.