In Press:

Send correspondence to: Entomologia Experimentalis et Applicata

Joseph M. Patt

Department of Entomology

Rutgers University

New Brunswick, NJ 08903 USA

Foraging success of parasitoid wasps on flowers: Interplay of insect morphology, floral architecture and searching behavior.

Joseph M. Patt, George C. Hamilton and James H. Lashomb Department of Entomology Rutgers University New Brunswick, NJ 08903, USA

Key words: searching behavior, biocontrol, parasitoids, Edovum puttleri, Pediobius foveolatus, intercropping, floral architecture, Anethum graveolens, Foeniculum vulgare, Coriander sativa.

ABSTRACT

Intercropping with flowering herbaceous plants increases parasitoid survivorship, fecundity and retention and pest suppression in agroecosystems. Few stu dies, however, have examined the compatibility of parasitoid morphology and foraging ability with floral architecture. This study shows that floral architecture influences the selection of floral host resources used to provide nutrients to parasitoids in cropping systems. Parasitoid foraging performance was evaluated using real and artificial flowers which varied in degree of nectar accessibility for two eulophid parasitoids, Edovum puttleri Grissell and Pediobius foveolatus Crawford. Comparisons were made of searching performance on artificial flowers with nectars that were either scented (made from 1:1 honey-water solution) or scentless (made from 1 M sucrose solution) and differences in head widths were compared with corolla apertures. Our results showed a disparity in the ability of E. puttleri and P. foveolatus to gain access to nectar from particular types of floral architectures. E. puttleri fed efficiently only from flowers with exposed nectaries while P. foveolatus foraged efficiently from flowers having either exposed nectaries or nectaries partially obstructed by petals and stamens. Neither wasp species could forage on flowers with cup- or tube-shaped corollas because their heads are wider than the floral apertures. E. puttleri's foraging performance decreased as nectar inaccessibility increased in the artificial flowers, while P. foveolatus' foraging performance was uniform among the different artificial flowers. This indicates that E. puttleri has less propensity to search small openings for nectar than does P. foveolatus. The foraging success of both E. puttleri and P. foveolatus on artificial flowers was lower when 1 M sucrose solution was used as an artificial nectar rather than honey-water solution, indicating that the wasps were stimulated and attracted by the nectar odor. Our systematic evaluation of floral architecture with respect to parasitoid foraging ability has enabled us to predict which types of flowers would serve as suitable floral host plants for parasitoids in the field. That is, only flowers with nectaries that are completely exposed would function as suitable floral host plants for E. puttleri, while P. foveolatus could forage on flowers with either exposed or partially exposed nectaries. Examples of potentially suitable floral hosts suggested from our study include dill (Anethum graveolens L.) and fennel (Foeniculum vulgare L.) for both E. puttleri and P. foveolatus and coriander (Coriandrum sativa L.) for P. foveolatus.

INTRODUCTION

Since critical nutrients for adult parasitoids are often scarce in large monocultures, provisioning of food resources via the interplanting of flowering herbaceous plants within the cropping system may be required for biological control programs to succeed (Wolcott, 1942; Clausen, 1956; Leius, 1967; Syme, 1975; Zandstra & Mootka, 1978; Altieri & Whitcomb, 1979; Altieri & Letourneau, 1982; van Emden, 1989; Andow, 1991; King & Olkowski, 1991; Grossman & Quarles, 1993). For example, parasitism of green pea aphids (Myzus persicae Sulzer) by the wasp Diaeretiella rapae McIntosh is enhanced when sweet alyssum (Lobularia maritima L.) is interplanted within the lettuce fields (W. Chaney pers. comm., Grossman & Quarles, 1993); predation of cabbage aphid Brevicoryne brassicae L. by syrphid larvae increases when broccoli is intercropped with flowering mustards (Kloen & Altieri, 1990); high rates of parasitism of codling moth (Carpocapsa pomonella L.) eggs by Trichogramma were reported in Russian apple orchards sown with dill (Anethum graveolens L.), buckwheat (Fagopyrum sp.), and mustard (Brassica sp.) (Zandstra & Mootka, 1978). These increases in efficacy were due apparently to the combined effects of increased survivorship, fecundity, retention and immigration (Altieri & Whitcomb, 1979; Altieri & Letourneau, 1982). While pollen and nectar support metabolism and gamete development, flowers can also provide mating sites, alternative larval hosts and prey (Leius, 1963; Altieri & Whitcomb, 1979; Toft 1983). Floral color and scent may attract parasitoids from a distance and promote immigration from areas lacking food resources (Leius, 1960; Wäckers, 1994).

Parasitoid wasps are a conspicuous element of the insect fauna inhabiting small fragrant flowers (Leius, 1960; van Emden, 1963; Hirose,1966; Judd, 1970; Proctor & Yeo, 1972; Kevan, 1973; Buntin, 1983; Lindsey, 1984; Bugg & Wilson, 1989; Maingay et al., 1991; Jervis et al., 1993). However, little is known about their floral preferences, floral foraging behavior, and attraction to floral color and odor (Shahjahan 1974; Nilsson, 1981; Jervis et al., 1993 & 1996; Wäckers, 1994). This is unfortunate because a particular wasp's behavioral and physical ability to manipulate floral parts to obtain nectar or pollen may limit its foraging range to only certain types of flowers (Faegri & van der Pijl, 1973; Nilsson, 1981; Jervis et al., 1993; Idris & Grafius, 1995). Therefore, evaluation of the foraging performance of a particular parasitoid wasp species on certain flowers is a critical requisite for determining which types of plants are suitable floral hosts for that wasp species (Syme, 1975; Jervis et al., 1993; Idris & Grafius, 1995). Studies have shown that this is true for anthophilous predatory insects as well (Gilbert, 1981; Haslett, 1989; Cowgill et al., 1993).

Currently, we cannot predict if a particular parasitoid's foraging capabilities and morphology are compatible with flowers having certain types of floral architectures; i.e., the arrangement of petals, stamens and other floral parts in relation to nectar gland location. As a step towards remedying this situation, we systematically evaluated the foraging performance of two parasitoid wasps, Edovum puttleri Grissell and Pediobius foveolatus (both Eulophidae), on a variety of real and artificial flowers with disparate floral architectures. We also compared the parasitoid's head dimensions with the widths of the corolla apertures. This permitted comprehensive evaluation of the ability of E. puttleri and P. foveolatus to obtain pollen and nectar from flowers with particular floral architectures.

E. puttleri and P. foveolatus were chosen for this study because: 1) Plant micro-architectural features (e.g., trichomes) can affect the searching behavior of small parasitoids (Rabb & Bradley, 1968; Brewer et al., 1983) and, therefore, the constraints imposed by floral architecture are more likely to be more apparent with small parasitoids; 2) They differ in size, E. puttleri is 1.5-2.5 mm long while P. foveolatus is 2.0-3.5 mm long, and differences in body size may translate into differential foraging performance on various floral architectures; and, 3) Both species are important biological control agents in New Jersey: E. puttleri suppresses Colorado potato beetle (Leptinotarsa decemlineata Say (Chrysomelidae)) in eggplant (Lashomb et al., 1987; Lashomb, 1989) while P. foveolatus suppresses Mexican Bean Beetle (Epilachna varivestis Mulsant (Coccinellidae)) in soybean (Angelet et al., 1968; Hooker & Barrows, 1989).

MATERIALS AND METHODS

Compatibility of wasp morphology and foraging behavior with floral architecture. All studies were conducted using female wasps. The plant species selected for evaluation have small and fragrant flowers arranged in inflorescences and are either known floral host plants of other chalcidoid wasps or are from genera containing known chalcidoid floral hosts (Proctor & Yeo, 1972; Faegri & van der Pijl, 1978; Kevan, 1973; Jervis, 1993). These flowers were grouped into one of five different types of floral architectures based on degree of nectary accessibility (Fig. 1): 1) Exposed nectaries on an umbel: parsnip, Pastinica sativa L. (Apiaceae); bupleurum, Bupleurum rotundifolia L. (Apiaceae); rue, Ruta graveolens L. (Rutaceae); fennel, Foeniculum vulgare L. (Apiaceae); parsley, Petroselinum crispum L. (Apiaceae); carrot, Daucus carota L. (Apiaceae); angelica, Angelica archangelica L. (Apiaceae); and, dill, Anethum graveolens L. (Apiaceae). These plants have fragrant flowers with glistening nectaries situated on the upper surface of inferior ovaries. The petals reflex-downward and away from the nectary while the stamen arch above it, so that the wasps are apparently not obstructed and can easily move across the petals and between the stamen to reach the nectar. 2) Exposed nectaries on a cyathium (all Euphorbiaceae): cypress spurge, Euphorbia cyparissius L.; spotted spurge, E. maculata L.; flowering spurge, E. maculata L.; and, snow-on-the-mountain, E. marginata Pursh.. In spurges, two bracts subtend the inflorescence, forming an open shallow bowl, and nectar is secreted by an exposed ring of five glands arising from the base of the inflorescence (Fig. 1). The stamen and style of the spurge flowers arch over the nectar glands and apparently do not hinder wasp access to the nectar. 3) Partially hidden nectaries on an umbel (all Apiaceae): ammi, Ammi majus L. ; coriander, Coriandrum sativa L.; and, culantro, Eryngium foetidum L.. In lateral view, the nectaries of these plants appear to be partially-obstructed by the bi-lobed petals which strongly reflex upwards and by the coiled filaments of the stamens. Because the wasps can see these nectaries they are not hidden per se, but access is made potentially more difficult by obstructions caused by the other floral parts. 4) Partially hidden nectaries in cup- or bowl-shaped flowers: sweet alyssum, Lobularia maritima L. (Brassicaceae); spearmint, Mentha spicata L. (Lamiaceae); chickweed, Stellaria media L. (Caryophyllaceae); sheperd's purse, Capsella bursa-pastoris L. (Brassicaceae). These flowers have nectaries that are recessed below the corolla aperture and are apparently obstructed laterally by the corolla. 5) Hidden nectaries in a capitulum (all Asteraceae): ageratum, Ageratum houstonianum Mill.; yarrow, Achillea millifolium L.; galansoga, Galansoga parviflora Cav.; and, chamomile, Matricaria chamomila L.. The nectaries of these plants are located at the base of the narrow tubular corollas of the disc flowers. We also evaluated the foraging behavior of each wasp on the extra-floral nectaries located on the stipules of snap beans (Phaseolus vulgaris L.(Fabaceae)) (Palmer, 1978). Because the nectar is exuded directly onto the stipule surface, these observations provided a baseline with respect to the wasps' foraging abilities on completely exposed surfaces. P. foveolatus was tested on selected flower species that were deemed representative of each type of floral architecture.

Figure 1. Diagrammatic representation in lateral view of the floral architectures on which E. puttleri and P. foveolatus were evaluated showing position of the nectar glands (in black) in relation to the other floral parts: 1) Umbels with exposed nectaries; 2) Cyanthia with exposed nectaries; 3) Umbels with partially hidden nectaries; 4) Cup-shaped flowers with partially hidden nectaries; 5) Capitula with hidden nectaries. Wasps are drawn to scale and are 3 mm long.

Prior to each test, two- to three-day old wasps were held in 4 l glass culture jars with watering stations but no food to stimulate hunger and maintained in incubators at 25.0° + 2.0°C and a L16:D8 photoperiod. To achieve satisfactory results, E. puttleri was held overnight while P. foveolatus was starved for two nights prior to each observation or test period. Flowers were cut just prior to each observation period from plants grown either in a greenhouse or adjacent research garden and the stems were immediately placed into 100 ml Erlenmyer flasks containing tap water. The plants were grown from seed in 25 cm diameter pots containing standard potting mix and fertilized weekly. Insecticidal soap was applied to the foliage, but not to the flowers, to control pest insects. During the summer months some of the plants were transplanted to a research garden adjacent to the greenhouse.

The wasps were transferred directly from the holding jars to the flowers which were illuminated with Reichartâ 150 W lamp with fiber optic arms. Wasp behavior was recorded with a Panasonic Digital 5000â color video camera equipped with a Computarâ 55 mm, f 2.8 telecentric lens, which provided enough magnification to observe mouthparts movements. The observations were conducted during the day between 10:00 and 15:00 h in a laboratory maintained at 25.0° + 2.0°C. Video tapes were examined to determine if E. puttleri and P. foveolatus performed the following key components of foraging behavior on each candidate floral host: 1) Feeding directly from nectary and anthers; 2) Pollen feeding from petals and other floral parts; 3) "Pollen grooming", in which the wasp feeds on pollen it gathers from its body surface; and, 4) Active exploratory movements and extensive antennation of floral parts. A dissection microscope fitted with an ocular micrometer was used to measure wasp head width and the length of the gap between the petals and the stamen or style of flowers selected as representative of floral architectures with partially exposed (coriander), partially hidden (alyssum and spearmint) and hidden nectaries (yarrow). Measurements were made of 50 E. puttleri and P.foveolatus heads and 25 flowers of each of species. Observations on artificial flowers. To determine if E. puttleri and P. foveolatus foraging success on different floral architectures was constrained by searching behavior, we examined their ability to locate nectar droplets positioned within a series of plastic flowers set within test arenas. Each test arena consisted of 16 flowers, arranged into an inner and outer ring of eight flowers each, glued with epoxy to the backside of a glass petri dish (Fig. 2a). The artificial flowers were similar in size to the flowers tested above and varied with respect to nectar accessibility in the following degrees: 1) exposed; 2) partially exposed; 3) partially hidden; and, 4) hidden (Fig. 2b).

A single test arena was constructed for each type of artificial flower. The artificial flowers were constructed from yellow concave (2.8 mm diameter) honeybee tags (Opalith-Plättchen, Chr. Graze KG, Stuttgart). The flowers with "exposed nectaries" were constructed by gluing the tags, with the concavity facing upwards to create a small well, into 1mm-deep depressions in the surface of the petri dish made with a glass bit-fitted drill tool. The flowers with the "partially exposed" nectaries were constructed by first melting a small hole with a heated needle into the center of individual tags and then gluing them concavity-side down onto the surface of the Petri dish. To construct the flowers with "partially hidden" and "hidden" nectaries, the disks were cut into pieces to create artificial petals. Each artificial flower was created by arranging four plastic petals into a circle and gluing them to the surface of a glass petri dish. The "corolla apertures" of the flowers with the "partially hidden" nectaries were 1.0-1.5 mm above the surface of the petri dish while that of the flowers with "hidden" nectaries were 2.0-2.5 mm above the surface

Figure 2. A. Diagrammatic representation of test arena as viewed from above showing arrangement of 16 artificial flowers in two circles. B. Cross sections of the four types of artificial flowers showing degree of nectar accessibility.

(Fig. 2b). A 1.0 ml droplet of 1:1 honey:water solution was added with a microsyringe to each flower to simulate nectar. In all but the artificial flowers with exposed nectaries, the nectar droplet was not visible to wasps until they discovered it. Prior to the start of each test, a starved wasp was placed onto a upright filter paper disk, supported by a wire frame, which was spotted with 1 M sucrose solution to induce localized searching behavior (Lewis & Takasu, 1990). After feeding for 60 s, the wasp was transferred to the test arena. Each wasp was tested only once and was observed for 300 s or until it either discovered a nectary or left the test arena. For each type of artificial flower, the time to nectar droplet discovery and the number of wasps that discovered nectar droplets were recorded (Number of wasps tested: E. puttleri n = 40; P. foveolatus n = 60). The artificial flowers were always tested sequentially with exposed nectaries presented first, partially exposed presented second, partially hidden tested third, and hidden nectaries presented last, with a full sequence of tests completed in 25-30 min. Only a single wasp species was tested on a given day and experiments were not performed if the wasps failed to become stimulated by the 60 s contact with 1 M sucrose solution. Role of floral scent in parasitoid searching behavior. Since scent can play a primary role in flower location by parasitoids (Leius, 1960; Shahjahan 1974; Lewis & Takasu,1990; Wäckers, 1994), we determined if E. puttleri and P. foveolatus were influenced by nectar odor by comparing wasp searching behavior in scented and unscented artificial flowers. The same artificial flowers arrays from the previous experiment were used. However, because sucrose solution does not induce a strong olfactory response in parasitoids (Lewis & Takasu, 1990), the nectar droplets were composed of 1 M sucrose solution instead of honey solution. The experiments were conducted as described above, with discovery time and number of wasps discovering flowers determined for each type of artificial flower. Statistical Analysis. For each wasp species, percentage of time foraging on nectar and total time on each flower was appropriately transformed (Snedecor & Cochran, 1978) and analyzed by analysis of variance (ANOVA) (SAS 1987). Comparisons of nectar droplet discovery time by artificial flower type (exposed, partially exposed, partially hidden and hidden) and nectar type (honey or sucrose solution) were made using the Kaplan-Meier Product-Limit Estimate (Statistica, StatSoft 1995), one type of survivorship analysis. For this analysis we substituted time to death with time to discovery of nectar droplets. The numbers of wasps that discovered nectar droplets in the artificial flowers was compared between artificial flower type and nectar droplet type by Contingency Table Analysis (Zar 1974).

RESULTS

Compatibility of wasp morphology and foraging behavior with floral architecture--E. puttleri and P. foveolatus quickly located and fed on the nectar produced by the exposed extra-floral nectar glands of snap bean, but varied in their ability to access nectar from the various flowers presented to them (Table 1). Both wasps quickly located the nectaries of all of the umbel and cyanthia with exposed nectar glands. While feeding they perched on the petals or sat directly on the exposed nectaries and typically fed on the umbel and spurge nectaries for periods lasting 300 s or more (Table 1). Nectar-feeding was frequently interrupted by antennation of the nectary and other floral parts, and by pollen-feeding either directly from the anthers or from other floral parts onto which pollen had fallen, or by grooming. Active exploratory movements and extensive antennation of floral parts resulted in retention times of the parasitoids on single florets that typically lasted in excess of 500 s (Table 1).

Edovum puttleri had great difficulty gaining access to the nectar in other umbelliferous flowers with partially exposed nectaries, such as coriander, culantro and ammi (Table 1). While E. puttleri vigorously attempted to separate the petals and stamen of ammi and coriander flowers to reach the nectaries, less than half successfully contacted the nectary. None of the E. puttleri tested were able to access the nectaries of culantro but were occasionally observed scavenging nectar residue left on coriander corollas as various species of foraging bees retracted their tongues while leaving the flowers. In contrast, P. foveolatus could insert its head between the petals and stamen of coriander and was able to feed on this flower's nectary (Table 1).

Figure 3. Comparison of the mean (+ s.e.) head width of E. puttleri and P. foveolatus with the mean (+ s.e.) length of the gaps between the petals and stamen (spearmint and alyssum) or petals and style (coriander and yarrow) of representative flowers with partially exposed (coriander), partially hidden (spearmint and alyssum) and hidden (yarrow) nectar glands. N = 50 E. puttleri and P. foveolatus and 25 of each flower.

Edovum puttleri and P. foveolatus could not forage efficiently on bowl- or cup-shaped flowers such as those of alyssum and spearmint (Table 1). This is because the head and thorax of each wasp species is wider than the gap between the petals and stamen of these flowers (Fig. 3) and cannot be extended into the corolla apertures to reach the nectaries. Typically the wasps rapidly crawled across several flowers and then either flew or sat and groomed on the corolla or pedicel and did not initiate foraging movements for the remainder of the observation. In a few observations E. puttleri and P. foveolatus fed from flowers whose petals that had been widely separated by foraging bees.

None of the four composites tested appeared to be suitable floral hosts for either wasp species, as the wasps moved quickly across the disc flowers to the rays and never probed the tubular disc flower corollas for nectar (Table 1). The heads of both E. puttleri and P. foveolatus are wider than the gap between the style and the corolla (Fig. 3), so they cannot probe into the flowers for nectar. While crawling on the disc flowers, pollen accumulated on their body surface and they commenced grooming after they reached the ray flowers. Although the wasps fed on the pollen as they groomed they frequently flew from the flowers immediately after grooming. Observations on artificial flowers--Edovum puttleri foraging performance differed among the various artificial flowers. In the exposed artificial flowers, E. puttleri tested discovered the honey nectar in 60 s or less (Table 2a & Fig. 4a). However, as the degree of nectar droplet inaccessibility increased, the time to nectar discovery increased (Fig. 4a). Significantly fewer (c2 = 27.08, P < 0.001) E. puttleri discovered the nectar droplets in the artificial flowers with hidden nectar than in the artificial flowers with more accessible nectar droplets (Table 2a). Pediobius foveolatus was more efficient in locating nectar droplets than E. puttleri, irregardless of nectar accessibility (Table 2b, Fig. 5a) .

Figure 4. Time to discovery of nectar droplets made from A) 50:50 honey-water solution; and, B)1 M sucrose solution in each type of artificial flower by E. puttleri. Comparisons of nectar droplet discovery time within each type of artificial flower was made using the Kaplan-Meier Product-Limit Estimate. N = 40 E. puttleri tested per each type of artificial flower.

Figure 5. Time to discovery of nectar droplets made from A)1 M sucrose solution; and, B) 1 M sucrose solution in each type of artificial flower by P. foveolatus. Comparisons of nectar droplet discovery time within each type of artificial flower was made using the Kaplan-Meier Product-Limit Estimate. N = 60 P. foveolatus tested per each type of artificial flower

both species was strongly affected by nectar scent. Significantly fewer wasps (E. puttleri: c2 = 12.33; P. foveolatus: c2 = 42.98, P < 0.001) discovered sucrose nectar droplets than honey droplets in the artificial flowers with less accessible nectar (Table 2), and sucrose nectar droplet discovery time also significantly increased (E. puttleri: c2 = 70.85; P. foveolatus: c2 = 22.56, P < 0.001) as nectar inaccessibility increased (Figs. 4b & 5b).

DISCUSSION

Our results showed a difference in the floral foraging ability between E. puttleri and P. foveolatus due to floral architecture. While both species foraged effectively on flowers with exposed nectaries, only P. foveolatus could efficiently forage on flowers with partially exposed nectaries. However, neither species could forage for nectar from flowers with partially hidden to hidden nectaries. Since their heads and thoraces are wider than the floral apertures of many small flowers and their mouthparts are very short, E. puttleri and P. foveolatus appear to be physically constrained from probing into narrow cup- or tube-shaped flowers. It is not surprising that small structures such as petals and stamen filaments can impede E. puttleri and P. foveolatus' access to nectar glands since the searching activities of other chalcids have been shown to be affected negatively by trichomes and other plant micro-architectural features (Rabb & Bradley, 1968; Brewer, et al., 1983). Because of its small size (< 3.5 mm) E. puttleri may not be strong enough to separate even the smallest floral parts of narrow cup- and tube-shaped flowers. This is suggested by the observations of E. puttleri and P. foveolatus foraging on the partially exposed nectaries of coriander flowers. On coriander P. foveolatus could access the nectaries by separating the petals and stamen with its head and legs but E. puttleri was not able to access the nectary, even though it attempted to penetrate the corolla by kicking vigorously and probing the petals and stamen. P. foveolatus' slightly larger size may confer enough strength to enable it to manipulate the floral parts of coriander and other flowers with partially hidden nectaries.

Another example where larger wasps are better able to handle flowers is provided by Idris & Grafius (1995) who observed the ichneumonid Diadegma insulare (Cresson) forcing apart the petal bases of Brassica kaber ((D. C.) Wheeler) to gain access to the nectaries located at the base of the corolla tube. They observed that D. insulare foraged best from flowers with relatively wide corolla apertures and showed a strong correlation between corolla aperture size and longevity and fecundity in D. insulare reared on various wildflowers in the lab. In addition, Jervis et al. (1993) noted that the head width and short length of the mouthparts restricted many parasitoid wasps to foraging only from flowers with exposed nectar glands, such as those in the Apiaceae. They found that only wasps with extremely narrow heads and bodies or with elongated mouthparts were able to gain access to nectar in flowers with tubular corollas.

Results of the artificial flower experiments indicate that E. puttleri and P. foveolatus differ in food searching efficiency. P. foveolatus was able to quickly discover the honey nectar droplets in all of the artificial flowers and only two out of 240 P. foveolatus tested failed to locate honey nectar droplets, showing that P. foveolatus can efficiently search small openings for scented nectar. In contrast E. puttleri displayed little propensity to search small openings for nectar. In the artificial flowers having inaccessible nectar, E. puttleri took longer to discover the nectar droplets than P. foveolatus and few E. puttleri discovered the nectar droplets in these artificial flowers. The searching performance of the wasps on both real and artificial flowers indicates that E. puttleri forages poorly on all flowers except those with exposed nectaries while P. foveolatus, because of its larger size and superior searching ability, can access nectar from flowers with both exposed and partially exposed nectaries.

When 1 M sucrose solution was used as an artificial nectar rather than honey-water solution, nectar discovery time increased while the number of wasps discovering nectar decreased, suggesting that the searching behavior of both parasitoid species is affected by nectar scent. These results suggest that nectar scent stimulates E. puttleri and P. foveolatus searching behavior and enables them to orient to the nectar within the artificial flowers.

Systematic evaluation of floral architecture with respect to parasitoid foraging ability has enabled us to predict that only flowers with exposed nectaries would function as suitable floral host plants for E. puttleri, while P. foveolatus can potentially exploit flowers with either exposed or partially exposed nectaries. Examples of suitable floral hosts from this study would include dill and fennel for both E. puttleri and P. foveolatus and coriander for P. foveolatus. Our findings indicate that only flowers whose floral architectures are compatible with a given insect's morphology and floral foraging ability can provide nutrients to that insect. This is in agreement with findings by Idris & Grafius (1995) and Jervis et al. (1993) whose studies provides other examples in which parasitoid foraging success is constrained by floral architecture and indicates that there exists a certain range of flowers from which parasitoids can forage.

Other factors, such as diurnal pattern of nectar secretion, nutritional value of nectar and pollen, flowering phenology and the floral host plant's cultural compatibility with the cropping system must also be carefully considered when choosing floral host plants for interplanting (Syme, 1975; Cowgill et al., 1993; Jervis et al., 1993). A consideration of these factors, along with further systematic evaluations of the floral foraging ability of parasitoids is needed to determine which flowers can provide nutrients to particular biological control agents within given crops.

ACKNOWLEDGMENTS

The authors wish to acknowledge the help of their undergraduate research assistants: William Merritt, Barbara Dove, Jamie Furneisen, Christine Makosky, Erol Sati, Nicole Synder, Elyse O'Grady and Amanda Dice. We are indebted to W. Joe Lewis of the USDA-ARS Lab in Tifton, Georgia for advice on the behavioral bioassays; Richard Trout, NJAES Statistician and Lisa Reed of the Rutgers Department of Entomology for assistance with statistical analysis; Dan Palmer and Robert Chianese of the New Jersey Department of Agriculture's Philip Alampi Beneficial Insect Laboratory for providing us with E. puttleri and P. foveolatus; and Angela Racz for editorial assistance. We are grateful to Robert Bugg, James Cane, Mark Jervis, Michael May, Tai Roulston, Blair Sampson and an anonymous reviewer for their helpful comments on the manuscript.

REFERENCES

Altieri, M. A. & W. H. Whitcomb, 1979. The potential use of weeds in manipulation of beneficial insects. HortScience 14: 12-18. Altieri, M. A. & D. K. Letourneau, 1982. Vegetation management and biological control in agroecosystems. Crop Protection 1: 405-430. Andow, D. A., 1991. Vegetational diversity and arthropod population response. Annual Review of Entomology 36: 561-86. Angalet, G. W., L. W. Coles, & J. A. Stewart, 1968. Two potential parasites of the Mexican bean beetle from India. Journal of Economic Entomology 61: 1073-1075. Brewer, G. J., E. L. Sorensen, & E. K. Horber, 1983. Trichomes and field resistance of Medicago species to the alfalfa seed chalcid (Hymenoptera: Eurytomidae). Environmental Entomology 12: 247-251. Bugg, R. L. & T. Wilson, 1989. Ammi visnaga (L.) Lamark (Apiaceae): Associated beneficial insects and implications for biological control, with emphasis on the bell-pepper agroecosystem. Biological Agriculture and Horticulture 6: 241-268. Buntin, L. A. 1983, The relationship of wild parsnip, Pastinaca sativa L. to parasitoid populations associated with soybean pests in central Iowa. Ph.D dissertation. Iowa State University. Ames. Clausen, C. P., 1956. Biological Control of Insect Pests in the Continental United States. U. S. Dept. of Agriculture Tech. Bull. No. 1139, U. S. Gov. Printing Office, Washington, D. C. Cowgill, S. E., S. D. Wratten, & N. W. Sotherton, 1993. The selective use of floral resources by the hoverfly Episyrphus balteus (Diptera: Syrphidae) on farmland. Annals of Applied Biology 122: 223-231. Faegri, K. & L. van der Pijl, 1979. The Principles of Pollination Ecology, 3rd. Ed. Pergamon Press, New York. Gilbert, F. S., 1981. Foraging ecology of hoverflies: morphology of the mouthparts in relation to feeding on nectar and pollen in some common urban species. Ecological Entomology 10: 385-392. Grossman, J. & W. Quarles, 1993. Strip intercropping for biological control. The IPM Practioner 15: 1-11. Haslett, J. R., 1989. Interpreting patterns of resource utilisation: randomness and selectivity in pollen feeding by adult hoverflies. Oecologia 78: 433-442. Hirose, Y., 1966. Parasitic Hymenoptera visiting the flowers of carrot planted in the truck crop field. Science Bulletin of the Faculty of Agriculture, Kyushu University 22: 217-223. Hooker, M. E. & E. M. Barrows, 1989. Clutch sizes and sex ratios in Pediobius foveolatus (Hymenoptera:Eulophidae), primary parasites of Epilachna varivestris (Coleoptera: Coccinellidae). Annals of the Entomological Society of America 82: 460-465. Idris, A. B. & E. Grafius, 1995. Wildflowers as nectar sources for Diadegma insulare (Hymnetoptera:Ichneumonidae), a parasitoid of diamondback moth (Lepidoptera: Yponomeutidae). Environmental Entomology 24: 1726-1735. Jervis, M. A., N. A. C. Kidd, M. G. Fitton, T. Huddleson, & H. A. Dawah, 1993. Flower- visiting by hymenopteran parasitoids. Journal of Natural History 27: 67-105. Jervis, M. A. and N. A. C. Kidd, 1996. Insect Natural Enemies. Chapman and Hall, London. Judd, W. W., 1970. Insects associated with wild carrot, Daucus carota L. in southern Ontario. Proceedings of the Entomological Society of Ontario 100: 176-181. Kevan, P. G., 1973. Parasitoid wasps as flower visitors in the Canadian high arctic. Sonderdruck aus Anzeiger fur Schadlingskunde, Pflanzen- und Umweltschutz XLVI. Jahrg., Heft 1 (1973), Seite 3-7. King, S., & W. Olkowski, 1991. Farmscaping and IPM. The IPM Practioner 13: 1-12. Kloen, H. & M. A. Altieri, 1990. Effect of mustard (Brassica hirta) as non-crop plant on competition and insect pests in broccoli (Brassica oleracea). Crop Protection 9: 90-96. Lashomb, J., 1989. Use of biological control measures in the intensive management of insect pests in New Jersey. American Journal of Alternative Agriculture 3: 77-83. Lashomb, J. H., D. Krainacker, R. K. Jansson, Y. S. Ng, & R. Chianese, 1987. Parasitism of Leptinotarsa decemlineata (Say) eggs by Edovum puttleri Grissell (Hymenoptera: Eulophidae): effects of host age, parasitoid age, and temperature. Canadian Entomologist 119: 75-82. Leius, K., 1960. Attractiveness of different foods and flowers to the adults of some hymenopterous parasites. Canandian Entomologist 92: 369-376. Leius, K., 1963. Effects of pollens on fecundity and longevity of adult Scambus buolianae (Htg.) (Hymenoptera:Ichneumonidae). Canadian Entomologist. 95: 202-207. Leius, K., 1967. Influence of wild flowers on parasitism of tent caterpillar and codling moth. Canadian Entomologist 99: 444-446. Lewis, W. J. & K. Takasu, 1990. Use of learned odours by a parasitic wasp in accordance with host and food needs. Nature 348: 635-636. Lindsey, A. H., 1984. Reproductive biology of the Apiaceae I: Floral visitors to Thaspium and Zizia and their importance in pollination. American Journal of Botany 71: 375-384. Maingay, H. M., R. L. Bugg, R. W. Carson & N. A. Davidson, 1991. Predatory and parasitic wasps (Hymenoptera) feeding at flowers of sweet fennel (Foeniculum vulgare Miller var. dulce Battandier (Trabut), Apiacea) and spearmint (Mentha spicata L. Lamiaceae) in Massachusetts. Biological Agriculture and Horticulture 7: 363-383. Nilsson, L. A., 1981. The pollination ecology of Listera ovata (Orchidaceae). Nordic Journal of Botany 1: 461-480. Palmer, D., 1978. Extrafloral nectaries as food sources for Pediobius foveolatus. Report to the New Jersey Department of Agriculture, Trenton. Proctor, M. & P. Yeo, 1972. pp. 134-135. The Pollination of Flowers. Taplinger Publishing Co., New York. Rabb, R. L. & J. R. Bradley, 1968. The influence of host plants on parasitism of eggs of the tobacco hornworm. Journal of Economic Entomology 61 ; 1249-1252. SAS, 1987. SAS Institute. Cary, NC. Shahjahan, M., 1974. Erigeron flowers as a food and attractive source for Peristenus pseudopallipes, a braconid parasitoid of the tarnished plant bug. Environmental Entomology 21: 401-408. Snedecor, G. W. & W. G. Cochran, 1978. Statistical Methods. Iowa University Press. Ames, IA. Statistica, 1995. StatSoft, Inc., Tulsa, OK. Syme, P. D., 1975. The effects of flowers on the longevity and fecundity of two native parasites of the European pine shoot moth in Ontario. Environmental Entomology 4: 337-346. Toft, C. A. 1983. Community patterns of nectivorous adult parasitoids (Diptera:Bombylidae) on their resources. Oecologia 57: 200-215. van Emden, H. F. 1963. Observations on the effects of flowers on the activity of parasitic Hymenoptera. Entomologist's Monthly Magazine 98: 265-270. van Emden, H. F. 1989. Plant diversity and natural enemy efficiency in agroecosystems. In Critical Issues in Biological Control. Intercept Ltd., Andover, UK. Wäckers, F. L. 1994. Multisensory Foraging by Hymenopterous Parasitoids. Doctoral Thesis, University of Wageningen, The Netherlands. Wolcott, G. N. 1942. The requirements of parasites for more than host. Science 96: 317-323. Zandstra, B. H. & P. S. Motooka, 1978. Beneficial effects of weeds in pest management- A review. PANS 24: 333-338. Zar, J. H., 1974. Biostatistical Analysis. Prentice-Hall, Inc. Englewood Cliffs, NJ.