HORTICULTURAL ENTOMOLOGY
Honey Bee (Hymenoptera: Apidae) Foraging in Response to Preconditioning with Onion Flower Scent Compounds ERIN M. SILVA,1,2 BILL B. DEAN,1,3
AND
LARRY K. HILLER4
J. Econ. Entomol. 96(5): 1510Ð1513 (2003)
ABSTRACT Onion (Allium cepa L.) seed production has long been plagued with yield problems because of lack of pollination by the honey bee, Apis mellifera L. To attempt to attract more pollinators to the onion seed production Þeld, honey bees were conditioned to associate onion ßoral odor components with a reward. Isolated nucleus hives of honey bees were fed 30% sucrose solutions scented with a 0.2% solution of onion ßoral odor compounds. After feeding on these solutions for 6 wk, bees were not found to prefer onion ßowers to two competing food sources, carrot and alfalfa ßowers, at the 5% signiÞcance level. However, there was an overall trend indicating a change in honey bee behavior, with fewer “trained” bees visiting alfalfa and carrot and more visiting onion. Thus, it may be possible to alter honey bee behavior with preconditioning but probably not to a degree that would be economically signiÞcant. KEY WORDS onion, Allium cepa, pollination, honey bees, honey bee behavior
FLOWERS ARE RECOGNIZED FOR their distinctive and often pleasant odors. Beyond their esthetic appeal, ßoral scents have a practical aspect; the volatile compounds comprising the odor of the ßower may act as semiochemicals and have speciÞc functions in insect pollination and host plant selection (Metcalf 1987). Honey bees recognize ßowers by three ßoral characteristics: color, shape, and odor (von Frisch 1967). They use these characteristics to distinguish those ßowers that provide a reward and to base their subsequent visits (von Frisch 1967). Of all the senses, odor learning occurs most rapidly, with scent memory providing highly reliable choice behavior of ⬇98% (Menzel et al. 1974) Several studies have been performed to determine bee behavior in terms of odor and conditioning response. In one such study, the behavioral discrimination of oilseed rape volatiles by the honey bee was measured (Pham-Delegue et al. 1993). They found that the honey bee seemed to distinguish between qualitative and quantitative differences in ßoral cues related to plant genotype and use these differences to select the most desirable nectar source. These data also demonstrated a hierarchy of preference among the compounds, with some chemicals initiating mix1 Department of Horticulture and Landscape Architecture, Washington State University Irrigated Agriculture Research and Extension Center, 24106 North Bunn Road, Prosser, WA 99350. 2 Corresponding author, current address: USDAÑARS, Vegetable Crops Research Unit, Department of Horticulture, University of Wisconsin-Madison, 1575 Linden Dr., Madison, WI 53706 (e-mail:
[email protected]). 3 Current address: General Manager, Walla Walla River Farms, 2010 East Melrose St., Walla Walla, WA 99362. 4 Department of Horticulture and Landscape Architecture, Washington State University, Pullman, WA 99164.
ture recognition more effectively compared with others. Once conditioned to a scent, bees will continue to visit that scent with the expectation of a reward. In a study of Majorana syriaca L., honey bees were conditioned to the scent of MajoranaÕs ßowers and leaves by providing a 70% sucrose reward (Beker et al. 1989). During the test, the conditioning device was replaced by a second identical device that contained the scents but not the reward. When a bee discovered a previously known scent at a certain site, it landed and extended its proboscis, indicating that she identiÞed the location of the food source on the basis of scent. After honey bees have been imprinted with a particular odor, their response to a different concentration of the same odor is still unclear. Vareschi (1971) found that bees conditioned to an odor did not respond to slight differences in the concentration of the odor. Pelz et al. (1997) discovered that odors at high concentrations support stronger associations compared with the lower concentrations but are not responded to by the bees as two distinctly different stimuli. Additionally, they found that honey bees could be trained to discriminate between concentrations of one odor, but could not discriminate between solutions of rewarded low odor concentrations and unrewarded high odor concentrations Pham-Delegue et al. (1993), however, saw that bee response to an odor after conditioning could only be elicited by a narrow range of concentrations above and below those that used for conditioning. Onion ßowers are notoriously unattractive to the honey bee. In a previous study, we found that unattractive onion lines produced signiÞcantly less nectar compared with the attractive lines (Silva and Dean
0022-0493/03/1510Ð1513$04.00/0 䉷 2003 Entomological Society of America
October 2003
SILVA ET AL.: BEE RESPONSE TO ONION FLOWER SCENT COMPOUNDS
2000). We also identiÞed the volatile chemicals that comprise the onion ßower odor (Silva 1998). Previous attempts at attracting bees to speciÞc crops with scented solutions have been met with varying success. Waller (1972) found that scented sucrose solutions applied to plots of alfalfa resulted in more visitation by honey bees to the treated compared with untreated plots. After these positive results, he experimented with the same technique on onions. Essential oils (1 ml/liter of citral and 1 ml/liter geraniol) were added to sucrose solutions to make a syrup that was sprayed on the onion ßowers with either a hand sprayer, a small power sprayer, or airplane. However, bee visitations to the onion crop were no greater compared with before; as Waller states, “the lack of effectiveness of attractant sprays for increasing bee visitation to onion Þelds was disappointing.” Despite WallerÕs disappointment, we decided to attempt a honey bee conditioning experiment using a unique approach. The previously described experiment used odors that were supposedly attractive to the honey bee but not necessarily characteristic of the onion ßower itself. We decided to design a conditioning experiment in which odors actually identiÞed from the onion ßower were used, with an end goal of attracting greater numbers of foraging bees to the onion crop versus two competing crops (carrot and alfalfa) after the bees were initially placed in the Þeld. Materials and Methods Bee Conditioning Experiment. On 5 May 1999, 12 nucleus hives of honey bees (⬇12,000 bees) were obtained for the experiment. Two locations in remote areas of wheat Þelds were selected: one for the treated bees and one for the control. Because bees concentrate their forage expeditions in a limited radius from their hives (Waddington et al. 1994), the selected areas for the hives were located at least 2 mi from each other and any other foraging possibilities, thus increasing the probability that the bees would only forage on the solutions given. Bees were given 18 liters of a 30% sucrose solution (the approximate concentration of onion ßower nectar) (Silva and Dean 2000) every 3 d. The control hives were given unscented solutions; the treated hives were given solutions scented with the compounds (total of 0.2% of scents) listed in Table 1 (Sigma, Milwaukee, WI). These compounds were 11 of the most prevalent chemicals comprising the onion ßower odor. The individual concentrations were based on percentages of each component as they occur in volatile identiÞcations of onion ßowers (Silva 1998). Solutions were changed every 3 d to keep the odor levels constant. Bees were fed the solutions for 6 wk so that a completely new generation of foragers only experienced with the onion scented solutions would exist in the hive. Field Experiment. In 1998, three plots (109.7 by 9.1 m) of carrots, onions, and alfalfa were divided into six smaller subplots (18.3 by 9.1 m) in a randomized complete block design. Carrot and alfalfa were planted with onion as “competing” crops because they
1511
Table 1. Composition of the scented solutions used to condition honey bees
Compound
Propanethiol Benzaldehyde Limonene Ocimene Dipropyl disulÞde 2-methyl 2pentenal Methyl heptanone Methyl benzoate Heptanal Octanal Nonanal
Amount identiÞed Percentage of Amount added from onion total scent to 11 of 30% ßoral scent compoundsb sucrose (l)c (g compound/ g tissue/h)a 28.73 47.62 4.16 15.57 109.30 2.89
8.4 13.9 1.2 4.6 32.0 0.8
168 278 24 92 640 16
2.60
0.8
16
9.39 64.49 2.87 10.37
2.7 18.9 0.8 3.0
54 378 16 60
a
From Silva 1998. Calculated as percent of total amount of identiÞed scents. Amount needed to make 0.2% scent solutions in the same percentage of total scent compounds. b c
are known to bloom concomitantly with onion and draw bees away from onion seed Þelds. Onion plots were planted on 4 August 1998. Carrot plots were planted on 28 August 1998. Alfalfa plots were planted on 23 March 1999. All plots were cultivated using standard cultivation practices. At the time of bloom of each of the three crops (6 July 1999), one set of treatment hives was brought onto the Þeld. Two hives of either the control or scent-fed colonies were taken each time and placed 3.1 m away from the end of the Þeld. Hives were left on the Þeld for 1 d before counts were made. Bee counts were taken by walking slowly up and down the Þeld, counting the number of bees on the outermost two rows of crop. After 3 d, a new set of two hives from the opposite treatment was placed on the Þeld. Statistical Analysis. Analysis of variance (ANOVA) was conducted with Statview software (Abacus Concepts 1992). The P value was considered signiÞcant at the 5% level. Results and Discussion Overall, little honey bee activity was seen on the experiment plots. This may have been a result of the 24-h delay between the time that the hives were placed in the Þeld and the time that the initial observations were made; the bees may have already found an alternative source of food a distance from the experimental plots. Of the bees that did forage on the plot, the greatest number of bees from both the treated and control hives foraged on the carrot plots (Fig. 1). Species by bee treatment interactions were signiÞcant (P ⫽ 0.0432); therefore, each factor was segregated and analyzed individually. Overall treatment effects showed a P value of 0.0595. Looking at each species individually, the trained bees did not visit the onions signiÞcantly more compared with the un-
1512
JOURNAL OF ECONOMIC ENTOMOLOGY
Fig. 1. Mean ⫾ SE of number of trained bees foraging on different crop species before and after “training” (preconditioning). Bee counts reported as calculated average of total bees found on all subplots over a period of 12 d. Bee counts are total bees found on the outermost two rows of crop while slowly walking alongside rows.
trained bees (n ⫽ 48; F ⫽ 0.215; df ⫽ 1, 46; P ⫽ 0.645). Bee treatments were also not signiÞcant for alfalfa (n ⫽ 48; F ⫽ 0.742; df ⫽ 1, 46; P ⫽ 0.394). The greatest bee treatment effects were seen on carrots, whereas less trained bees foraged on the ßowers (n ⫽ 48; F ⫽ 3.538; df ⫽ 1, 46; P ⫽ 0.066). Looking at bee activity overall, it seemed that an extended period of feeding on scented sucrose solution led to an alteration in bee foraging behavior. Naõ¨ve foragers that had only been exposed to an unscented food source showed a great preference for carrot ßowers over alfalfa and onion. When fed the onion-scented sucrose solution, the degree of this preference declined (Fig. 1). The lack of increase in the level of foraging on the onion ßowers may have been caused by the nature of the scent mixtures. Although most of the major onion ßoral scents were present, the scented solutions were not an exact replication of the actual onion ßoral odor. Dipropyl disulÞde, the compound very characteristic of onion odor, is unique to Allium ßowers. If bees locate food sources on the basis of individual scents, this odor should have allowed the bees to easily Þnd onion as a food source. However, if bees use a characteristic odor pattern to Þnd their food sources (i.e., if the relative concentrations of individual scents comprising the entire odor is important), the scented solutions used for conditioning may not have given the bees the appropriate directions for Þnding onions as a food source. Because their foraging behavior did seem to be altered, the honey bees may have been searching for the food source to which they were accustomed. In addition, as stated above, it is still uncertain if, on conditioning, honey bees discern between different concentrations of the same odor. Further preconditioning studies should be done using a more precise representation of the onion ßower scent. Without a very signiÞcant increase in foraging behavior on the target crop, the economics of preconditioning would not prove to be beneÞcial to the seed producer. The preconditioning process was rather complicated and time-consuming and thus would be costly to the growers. Instead of trying to
Vol. 96, no. 5
train honey bees to forage on a crop that is not energetically rewarding, it would be better for plant breeders to improve the food source itself. It has been shown that onions do not provide enough nectar to make the crop proÞtable and thus attractive to honey bees. The importance of nectar ßow rate in the assessment of crop proÞtability to honey bees has been shown by several independent research groups (Varju and Nunez 1991, Greggers and Menzel 1993, Wainselboim et al. 2003). Thus, it behooves us to Þrst understand the physiology of onion nectar production and attempt to improve it through breeding methods. As Waller (1972) stated after his disappointing attempts with attractant sprays, “without a suitable reward from the target crop, bringing additional honey bees to a Þeld with an attractant spray is not likely to result in a permanent increase in pollinator activity. Therefore, improving the reward should be the ultimate goal in achieving a permanent solution to the problem of poor pollinator activity.” Acknowledgments We thank Robert Maxwell of Seminis Seed Company for providing onion and carrot seed. This project was supported by the American Seed Trade Association, the Columbia Basin Vegetable Seed Association, and by a USDA-CSREES Special Grant.
References Cited Abacus Concepts. 1992. Statview userÕs guide for the Macintosh. Abacus Concepts, Inc., Berkeley, CA. Beker, R., A. Dafni, D. Eisikowitch, and U. Ravid. 1989. Volatiles of two chemotypes of Majorana syriaca L. (Labiatae) as olfactory cues for the honey bee. Oecologia (Heidelb.). 79: 446 Ð 451. Greggers, U., and R. Menzel. 1993. Memory dynamics and foraging strategies of honeybees. Behav. Ecol. Sociobiol. 32: 17Ð29. Menzel, R., J. Erber, and T. Masuhr. 1974. Learning and memory in the honey bee, pp. 195Ð227. In L. BartonBrowne (ed.), Experimental analysis of insect behavior. Springer, Berlin, Germany. Metcalf, R. L. 1987. Plant volatiles as insect attractants. Crit. Rev. Plant Sci. 5: 251Ð301. Pelz, C., B. Gerber, and R. Menzel. 1997. Odorant intensity as a determinant for olfactory conditioning in honeybees: roles in discrimination, overshadowing and memory consolidation. J. Exp. Biol. 200: 837Ð 847. Pham-Delegue, M. H., O. Bailez, M. M. Blight, C. Masson, A. L. Picard-Nizori, and L. J. Wadhams. 1993. Behavioral discrimination of oilseed rape volatiles by the honey bee Apis mellifera L. Chem. Senses. 18: 483Ð 494. Silva, E. M. 1998. The effect of onion (Allium cepa L.) ßower characteristics on its pollination by the honey bee (Apis mellifera L.). MS thesis, Washington State University, Pullman, WA. Silva, E. M., and B. B. Dean. 2000. Effect of nectar composition and nectar concentration on honey bee (Hymenoptera: Apidae) visitations to hybrid onion ßowers. J. Econ. Entomol. 93: 1216 Ð1221. Vareschi, E. 1971. Duftunterscheidung bei der honigbiene. Einzelzell Ableitungen und Verhaltensreaktioner. Z. Vgl. Physiol. 75: 143Ð173.
October 2003
SILVA ET AL.: BEE RESPONSE TO ONION FLOWER SCENT COMPOUNDS
Varju, D., and J. A. Nunez. 1991. What do foraging honeybees optimize? J. Comp. Physiol. A Sens Neural Behav. Physiol 169: 729 Ð736. von Frisch, K. 1967. The dance language and orientation of bees. Harvard University Press, Cambridge, MA. Waddington, K. D., P. K. Visscher, T. J. Herbert, and M. Raviret Richter. 1994. Comparisons of forager distributions from matched honey bee colonies in suburban environments. Behav. Ecol. Sociobiol. 35: 423Ð 429. Wainselboim, A. J., F. Roces, and W. M. Farina. 2003. Assessment of food source proÞtability in honeybees (Apis
1513
mellifera): how does disturbance of foraging activity affect tropallactic behavior? J. Comp. Physiol. A Sens. Neural Behav. Physiol. 189: 39 Ð 45. Waller, G. D. 1972. Chemical differences between nectar of onions and competing plant species and probable effects on the attractiveness to pollinators. PhD dissertation, Utah State University, Logan, UT. Received for publication 11 February 2003; accepted 29 June 2003.
