Journal of Chemical Ecology, Vol. 20, No. 12, 1994

DISCRIMINATION OF OILSEED RAPE VOLATILES BY HONEY BEE: NOVEL COMBINED GAS

CHROMATOGRAPHIC-ELECTROPHYSIOLOGICAL BEHAVIORAL ASSAY

*To whom

L.J. W A D H A M S , I M . M . B L I G H T , t V . K E R G U E L E N , -~

M, LE M I ~ T A Y E R , 2 F. M A R I O N - P O L L , 3 C. M A S S O N , 2

M . H . P H A M - D E L I ~ G U E , 2'* and C . M . W O O D C O C K I

*Department of Bioh~gical and Ecological Clwmistry Rothamsted Experimental Station, hlstitute of Arable Crops Research

Harpenden, Hertfordshire AL5 2JQ, U.K.

~-Lztboratoire de Neurohiologie Compart;e des hlvertObr~;s INRA-CNRS (LIRA 1190)

BP 23. 91440 Bures-sur-Yvette, France.

~Station de Phytopharmaeie. INRA R~ute de St. Cyr, 78026 Versailles Cedex, France

(Received May 16, 1994: accepted August 4, 19941

Abstract--A n~wel technique for the simultaneous monitoring of etectroan- tennogram (EAG) and conditioned proboscis extension (CPE'~ responses of honey bees to the effluent from a gas chromatograph (GC) was developed to locate biologically active components in blends of plant volatiles and to inves- tigate odor recognition at the peripheral and behavioral levels. A six-com- ponent mixture, comprising compounds previously identified as oilseed rape floral volatiles, was used as the stimulus. Standard CPE and EAG recordings were done as a reference. EAG responses were elicited from unconditioned bees by all the components presented either in the coupled or the standard mode, Conditioned bees gave larger EAG responses than unconditioned bees, suggesting that antennal sensitivity is enhanced by conditioning. At the behav- ioral level, in both the standard and the coupled modes, only conditioned bees showed the proboscis extension response, with the majority of individuals responding to linalool, 2-phenylethanol. and benzyl alcohol,

Key Words--Honey bee, Apis mellifera, Hymenoptera, Apidae, gas chro- matography, electroantennogvam, conditioned proboscis extension, olfactory discrimination.

correspondence should be addressed.

3221

0098 -0331/94/1200 3221507.0~}/O ,c~ 1994 Plenum Publishing Corpt}ratlt~f

3 2 2 2 WADHAMS ET AL.

INTRODUCTION

Volatile semiochemicals have an important role in the chemical ecology of the honey bee and facilitate flower recognition, thereby increasing foraging el~- ciency. Chemical analyses of volatiles from various plant species have shown the complexity of floral odors. Tens to hundreds of components have been reported, e.g., 150 in orchid flowers (Borg-Karlson and Tengr, 1986), 80 in sunflower (Etirvant et al., 1984), Moreover, the quality and quantity of these blends can fluctuate according to various factors such as plant phenology [e,g,, cotton (Hedin, 1976); and sunflower (Pham-Del~gue et al., 1988)] or to pedo- climatic conditions (Robacker et al., 1982). It is still not known how honey bees process the chemical information and adapt their behavior in response to such complex and variable plant volatile blends.

Odor recognition in the honey bee has been investigated using a number of approaches. Behavioral experiments with free-flying bees demonstrated con- ditioning processes in food source recognition (Frisch, 1919) and showed the importance of olfactory rather than visual cues (Kriston, 1973). In addition, foragers were shown to establish preferences between odorants (Koltermann, 1969). With restrained bees, the conditioned proboscis extension response (CPE) devised by Frings (1944) was used to investigate learning processes (Bitterman et al., 1983) and their neural bases (Menzel et al., 1974; Erber, 1981; Menzel, 1990). Based on this bioassay, Vareschi (1971) made an extensive study of qualitative behavioral discrimination abilities among pairs of 28 odorants, and found that in 95.5% of cases, bees were able to discriminate between learned and unlearned odors. Getz and Smith (1987) demonstrated the quantitative dis- crimination abilities of bees by testing binary mixtures of the same compounds combined in different proportions.

Tentative work relating behavioral responses to peripheral detection pro- cesses has also been done. Thus, Vareschi (1971) established seven reaction groups from single-cell recordings and found that components from different groups were more systematically discriminated in the behavioral assays than components from the same reaction group. More recently, Ackers and Getz (1992) repeated this work on six components belonging to three reaction groups defined by Vareschi (1971) and found a different categorization of the receptor types.

In more applied studies aimed at improving plant pollination, chemical, behavioral, and electrophysiological techniques were used to identify potentially attractive plant volatiles. Thus, Waller et al. (1974) showed that free-flying honey bees trained to the scent of alfalfa flowers used components predominant in the training scent as cues. An active fraction cueing the recognition of a complex floral sunflower extract has also been found (Pham-Del~gue et al,, 1986). These behavioral studies were complemented by recording the antennal

VOLATILES DISCRIMINATION BY HONEY BEE 3223

responses of honey bees to floral volatiles, using a coupled gas chromatography (GC)-electroantennogram (EAG) detector system. In a whole-flower extract containing more than 100 components, only 24 compounds elicited EAG responses from more than 80% of the bees (Thi6ry et al., 1990) and of these compounds, only six had previously been found to be behaviorally active.

Electrophysiological recordings have been used widely to locate biologi- cally active compounds in complex natural product extracts (Am et al., 1975; Wadhams et al., 1982; Guerin et al., 1983; Thi6ry et al., 1990; Henning and Teuber, 1992). However, in many studies, attempts to correlate the activity of olfactory stimuli at the antennal level with the behavioral responses have been confounded because these phenomena have been investigated separately. To locate biologically active components in blends of plant volatites and to inves- tigate odor recognition at the sensory and behavioral levels, we have developed a novel technique for the simultaneous monitoring of EAG and CPE responses of honey bees to the effluent from a GC column. The CPE response, which occurs naturally when a forager visits a food source, is thought to indicate behavioral recognition of complex olfactory information by the bee (Mauels- hagen and Greggers, 1993). Since learning processes play an important role in floral visits by honey bees, it was decided that a bioassay based on such learning abilities should be used. Moreover, the use of restrained bees enabled the simul- taneous recording of behavioral and antennal responses to the same range of stimuli. A six-component mixture, comprising components previously identified as oilseed rape floral volatiles (Tollsten and Bergstr6m, 1988; Blight et al., 1992), was used as the stimulus to evaluate the technique. Standard CPE and EAG recordings were done as a reference.

METHODS AND MATERIALS

Chemicals

The chemicals used were linalool (97%), 2-phenylethanol (99 + %), methyl salicylate (99 + %), benzyl alcohol (99 + %), (E)-2-hexenal (99%), and 1-octen- 3-ot (98%). Solutions of the individual components, and mixtures containing equal amounts of all six components, were made up in hexane (HPLC grade, Fisons, purified).

Insects

Although all kinds of bees can be used in this bioassay, we worked with Italian bees because of their ability to endure caged confinement. Emerging Italian worker bees, Apis mellifera ligustica, were collected from outdoor hives and caged in groups of approximately 50 individuals. They were fed sugar,

3 2 2 4 WADHAMS ET AL.

pollen, and water ad libitum and maintained in an incubator at 33°C, 50% relative humidity, until required.

Bees 14-16 days old, were starved for at least 3 hr and mounted individ- ually in glass holders in such a way that their mouthparts and antennae remained free. For electrophysiological recordings, one antenna was fixed to the holder with water-based correction fluid (Tipp-Ex).

Coupled GC-EAG-CPE Recording

Bees tested in the coupled procedure were either unconditioned or were conditioned to the six-component mixture. The conditioning procedure (Pham- Del~gue et al., 1993) was carried out independently from the coupled system. Before odor stimulation, bees were habituated to mechanical stimulation by placing them for 15 sec in an airflow of 50 ml/sec, which was delivered through a l-cm-diameter glass tube placed 1 cm from the head of the bee. The stimulus delivery system utilized a disposable Pasteur pipet cartridge. The conditioning mixture (I ~g of each component in 10/~1 hexane) was applied to a 20 x 4-ram filter paper strip, the solvent was allowed to evaporate, and the filter paper was inserted into the pipet. Vapor from the cartridge was delivered over a 6-sec period into the airstream passing continuously over the bee by means of a second airstream (2.5 ml/sec) controlled by a solenoid. Fresh cartridges were prepared immediately prior to each stimulation. Three seconds after the beginning of the odor stimulation, the free antenna was contacted with a 30% sucrose solution and the proboscis extension was rewarded by offering a drop of the sugar solu- tion. Five conditioning trials were done for each bee, with 15-min intervals between trials. ARer the first trial, the conditioned response occurring during the first 3 sec of the odor stimulation was noted. A control trial, performed by stimulating the insect with hexane (10 p.I), was used to check whether the odor stimulus had been learned specifically or if the bee responded to other stimuli (solvent, airflow). Only bees that gave the conditioned response at least once during the conditioning trials and that did not respond to the control stimulation were used for the coupled GC-EAG-CPE recordings.

GC Conditions. Coupled GC-EAG-CPE recordings were obtained using the GC-EAG system described previously (Blight et al., 1979). Samples (six- component mixture, 0 .5/ag of each component) were separated using an A1-93 gas chromatograph equipped with a flame ionization detector (FID) and a cold on-column injector fitted with a 30-m x 0.53-mm-ID HP-1 bonded phase fused silica capillary column. The oven temperature was maintained at 40°C for 1 min and then programmed at 10°C/min to 250°C. The cartier gas was hydrogen. The effluent from the GC column was split approximately equally, with one part going to the FID of the GC and the remainder directed into the airflow passing continuously over the bee.

VOLAT1LES DISCRIMINATION BY HONEY BEE 3225

Electrophysiological Recordings. EAG responses were obtained using Ag- AgCI glass electrodes filled with saline solution [composition as in Maddrell (1969) but without the glucose]. The recording electrode was placed in a pre- punctured hole in the tip of the fixed antenna and the indifferent electrode was inserted into a small cut in the scape. Responses of conditioned and uncondi- tioned individuals were compared using a t test.

Proboscis Extension Recordings, CPE responses to the et~uent from the GC column were monitored visually and were recorded simultaneously with the EAG responses. FID, EAG, and CPE responses were stored on tape (Racal Store 4). Recordings were sampled on a microcomputer with a Data Translation card (DT2821; 12 bits precision: 200 Hz/channet) and analyzed with software developed initially for spike analysis (Marion-Poll and Tobin, 1992).

Standard Proboscis Extension Testing. At the end of the GC-EAG-CPE recordings, the bees were tested (6 sec unrewarded odor stimulation) with the individual components (1 t~g) presented in a random order at 15-min intervals, and with the original conditioning mixture. Bees from the conditioned group that did not give the CPE response to the conditioning stimulus at the end of the experiment was discarded from data treatment.

In this experiment, the responses from six conditioned and eight uncondi- tioned bees were recorded.

Standard Behavioral and Electrophysiological Recordings

In parallel with the coupled recordings, responses of bees were assessed using standard behavioral and electrophysiological procedures.

Standard CPE Assay. The conditioning procedure was as described pre- viously, After conditioning to the mixture (1 p,g of each component), bees were tested with the individual components (1 ~tg) presented randomly and with the conditioning mixture.

Electrophysiotogical Recordings. EAG recordings were obtained using the technique described for the coupled experiment. Dose-response curves were obtained for each component over a concentration range of 1-100 #g. Electro- physiological preparations were stimulated using the delivery system employed for the conditioned proboscis extension assay, except that the stimulus duration was 2 sec. All samples were presented twice to each preparation (total six bees) at intervals of 2 min. The EAG signals were amplified and recorded by standard methods (Wadhams et al., 1982).

RESULTS AND DISCUSSION

When bees were tested with the six-component mixture using the coupled GC-EAG-CPE technique, EAG responses were elicited by all the GC peaks

3226

J

A+B

F

D

J L

WADHAMS ET AL.

O

5 0 s

I • ...........

Fto. 1. Coupled GC-EAG-CPE recording on a conditioned honey bee using a six-com- ponent mixture (0.5 p.g of each component injected onto GC column). Upper trace: FID response. Compounds are: A, linalool; B, 2-phenylethanol; C, methyl salicylate; D, benzyl alcohol; E, (E)-2-hexenal; F, l-octen-3-ol. Middle trace: Simultaneous EAG response. Lower trace: Simultaneous CPE response.

(Figure 1). Under the GC conditions employed in this study, linalool and 2-phenyl ethanol were only partly resolved. EAG responses given by condi- tioned bees were larger than those elicited by the same stimulus concentration from unconditioned bees (Figure 2), although these differences were not signif- icant. However, the means of the total EAG responses to test compounds were significantly different [conditioned (X + SE) = 0.37 _+_ 0.04 mV; unconditioned = 0.23 + 0.02 mV; P < 0.05]. This suggests that antennal sensitivity is enhanced by conditioning. Similar results have been found with bees conditioned to violet and fennel odors (De Jong and Pham-Del~gue, 1991), and changes in the EAG sensitivity of the parasitic wasp Leptopilina heterotoma were observed after an oviposition experience (Vet et al., 1990). EAG responses obtained with unconditioned bees in the standard procedure showed all six compounds to have significant activity, with linalool, 2-phenylethanol, methyl salicylate, and 1-octen-3-ol eliciting the greatest responses, particularly at high stimulus con- centrations (Figure 3).

VOLATILES DISCRIMINATION BY HONEY BEE 3227

EAG (mY)

0.6

0.5

04 +

0 0 0 i 0

A+B C D E F m e a n s

FiG. 2. EAG responses (+SE) to compounds A-F (see Figure l) in the coupled GC- EAG-CPE assay: open columns, conditioned bees: shaded columns, unconditioned bees. *Means are significantly different at P < 0.05.

At the behavioral level, no proboscis extension response was obtained from any of the unconditioned bees, either in the coupled mode or when the com- pounds were presented in the standard CPE assay following GC stimulation. Of the six conditioned bees, only one (No. 5) did not respond to any of the indi- vidual compounds, presented either in the coupled mode or in the following standard CPE assay, although a response was still obtained to the conditioning mixture at the end of the testing protocol. The other five conditioned bees showed a close correlation between the CPE response to the components pre- sented either in the standard CPE assay or in the coupled GC-EAG-CPE mode (Table 1). In the coupled mode, all five bees gave a response to the linalool/2- phenylethanol peak, two responded to benzyl alcohol, and one responded to methyl salicylate. When presented separately in the standard mode, linalool elicited a response in all five bees. Three of them also responded to 2-phenylethanol. Two of the bees (No. 1 and 2) showed different responses to benzyt alcohol in the coupled and standard CPE assays. This is consistent with the fact that in a previous study using the same mixture, this compound was poorly discriminated by the honey bees (Pham-Del~gue et al., 1993). Overall, there was a good correlation between the responses of individual bees to the different components with the two stimulation techniques.

One hundred sixty-eight bees were subjected to the conditioning procedure prior to use in the standard CPE assay. Of these, 25% did not show any con- ditioned response during the conditioning trials, 28% responded to the control trial, and 24% did not respond to the conditioning mixture at the end of the testing. Therefore, the effectiveness of the conditioning in this study was poor, with only 23 % of the bees being successfully conditioned to the six-component

3228

2'

1,5,

Benzyl alcohol

:~ 1,5

W A D H A M S ET AL.

Linslool

-6 -5 -4 -6 -5 -4

Stimulus conc (Log g) Stimulus conc (Log g)

1,5

<~ 0,5 ud

1-Octer~3-ol (E)-2-Hexensl

:~ 1,5,

,..~ 0,5

-6 -5 -4 -6 -5 -4

Stimulus conc (Log g) Stimulus conc (Log g]

~ 1 , 5

c o ,

o,s

2-Phenylethsnol Methyl salicylste

i

=~ 1,5 ¸ ¢:

~0,5 , f I

i

-6 -5 -4 -6 -5 -4

Stimulus conc {Log 9) Stimulus conc {Log g)

FIG. 3. EAG dose-response data for unconditioned bees stimulated with test compounds (see Figure 1) at concentrations ranging from 1 to 100 ~g (means + SE of six prepa- rations. Responses to the solvent blank are shown in the lower right hand comer of each graph).

VOLATILES DISCRIMINATION BY HONEY BEE 3229

TABLE 1. C P E RESPONSES (CROSSES) OF SIX CONDITIONED HONEY BEES TO

COMPOUNDS A-F" tN COUPLED B1OASSAY FOLLOWED BY STANDARD C P E ASSAY

Responses to components

Coupled Standard

Bees A + B C D E F A B C D E F

l × ×

2 X X

3 x 4 x 5 6 x x Total 5 I 2

X

X X

X X

X X

X X

0 0 5 3 I 2 0 0

"See Figure 1.

mixture, Of the 39 conditioned bees, nine did not respond to any of the com- pounds when they were presented individually and hence did not recognize the mixture on the basis of its individual constituents. The remaining 30 bees all showed CPE responses to one or more of the components, with the majority (10 individuals) responding to two. Linalool, methyl salicylate, and 2-phenylethanol were the most active, eliciting responses from 25, 17, and 13 bees, respectively, Despite the lower conditioning levels obtained in this study compared to those observed by Pham-Del~gue et al. (1993) using the same mixture, overall response profiles of the conditioned bees were similar, with the six components being ranked in the same order of activity. These results also correlate well with the coupled GC-EAG-CPE data, with the linalool/phenyl- ethanol peak eliciting the highest responses. Differences in the proportion of bees responding to methyl salicylate and benzyl alcohol probably relate to the small sample size.

Thus, consistent with data obtained from standard CPE procedures, the application of the coupled method confirms the occurrence of a hierarchy within the constitutive components of the conditioning mixture, with some components eliciting mixture recognition more effectively than others (Pham-Del~gue et al., 1993). Moreover, presentation of the components via GC separation ensures that they are in the proportions occurring in natural plant volatile blends and also allows on-line purification, which, within the limits of column separation, permits avoiding interference from chemical impurities+ However, as with the standard CPE technique, this mode of stimulation provides a sequential pre- sentation of the components that does not allow evaluation of possible compo-

3230 WADHAMS ET AL.

nen t i n t e rac t ions ( e . g . , s y n e r g i s t i c / a n t a g o n i s t i c e f fec ts ) . T o c o m p l e m e n t th i s

work , fu r the r e x p e r i m e n t s wi th g r o u p s o f c o m p o n e n t s will be done .

In c o n c l u s i o n , th i s nove l c o u p l e d s y s t e m p r o v i d e s a u se fu l tool for the

i nves t i ga t i on and c h a r a c t e r i z a t i o n o f key c o m p o u n d s i n v o l v e d in m i x t u r e rec-

o g n i t i o n by h o n e y b e e s . It will a l l ow the r e l a t i onsh i p b e t w e e n o l f ac to ry d e t e c t i o n

and b e h a v i o r a l l y sa l i en t c u e s to be a s s e s s e d and the effect o f l e a r n i n g on o d o r

d i s c r i m i n a t i o n at bo th s e n s o r y and b e h a v i o r a l l eve l s to be d e t e r m i n e d .

Acknowledgments--The cooper'alive work between England and France was cofunded by the British Council and the French Minist~re de la Recherche et de la Technologie (Alliance Program).

REFERENCES

ACKERS, R.P., and GETZ, W.M. 1992. A test identified response classes among olfactory receptor neurons in the honey-bee worker. Chem. Senses 17(2): 191-209.

ARN, H., ST,h.DLER, E., and RAUSCHER, S. 1975. The electroantennograpbic detector--a selective and sensitive tool in the gas chromatographic analysis of insect pheromones. Z. Naturforsch. 30C:722-725.

BITTERMAN, M.E,, MENZEL, R., FIETZ, A., and SCHAFER, S, 1983. Classical conditioning of proboscis extension in honeybees (Apis nlellifera). J. Comp. Psycho1. 97(2): 107-119.

BLIGHT, M.M., HICK, A.J., PICKETT, J.A., SMART, L.E., WADHAMS, LJ . , and WOODCOCK, C.M. 1992. Volatile plant metabolites involved in host plant recognition by the cabbage seed weevil, Ceutorhynchus assimilis Payk, pp. 129-130, S.BJ. Menken, J.H. Visser, and P. Harrewij (eds). Proceedings, 8th International Symposium Insect-Plant Relationships. Kluwer Academic Publishers, Wageningen.

BORG-KARLSON, A.K., and TENGO, J. 1986. Odor mimetism? Key substances in Opho's lutea- atutrena pollination relationship (Orchidaceae: AndrenidaeL J. Chem. EcoL 12(9): 1927-1941.

DE JONG, R., and PHAM-DELt:GUE, M.H. 1991. Electroantennogram responses related to olfactory conditioning in the honey bee (Apis mellifera ligustica), J. h~sect PhysioL 37(4):319-324.

ERBER, J. 1981. Neural correlates of learning in the honeybee. TINS 4:270-273, ETIf~VANT, P.X., AZAR, M., PHAM-DELI~GUE, M H . , and MASSON, C. 1984. Isolation and identi-

fication of volatile constituents of sunflowers (Helianthus annuus L.). J. Agric. Food Chem. 32:503 -509.

FRINGS, H. 1944. The loci of olfactory end-organs in the honey-bee, Apis mellifera capensis Linn. J. Exp. Zool. 97:123-134.

FRISCH, K. VON 1919. Uber den Geruchsin der Bienen und seine bliitenbiologische Bedeutung. Z~oL Jahrb. 97:1-238.

GETZ, W.M., and SMITH. K.B. 1987. Olfactory sensitivity and discrimination of mixtures in honey- bee Apis mellifera. J. Comp, Physiol. A(160):239-245.

GUERm, P., STADLER, E., and BUSER, H.R. 1983. Identification of host-plant attractants for the carrot fly Psila rosae. J. Chem. EcoL 9(7):843-861.

HEDIN, P,A. 1976. Seasonal variations in the emissions of volatiles by cotton plants growing in the field. Environ. Entomol. 5:1234-1238.

HENNING, J.A., and TEUBER, L,R. 1992. Combined gas chromatography-electroantennograna char- acterization of alfalfa floral volatiles recognized by honey bee (Hymenoptera: Apidae). J. Econ. Entomol. 85(1):226-232.

KOLTERMANN, R. 1969. Learning and forgetting processes in the honey bee as revealed through scent training. Z'. Vergl. Physiol. 63(3):310-334.

VOLATILES DISCRIMINATION BY HONEY BEE 323 1

KRISTON, I. 1973, Die Bewertung yon Duff- und Farbsignalen als Orientierungshilfen an der Fut- terquelle durch Apis mellifera. L. J, Comp. Physiol. 84:77-94.

MaDDRELL, S.H.P., 1969. Secretion by the malpighian tubules of Rhodinius. The movement of ions and water. J. Exp. Biol. 51:71-97.

MARION-POLL, F., and TOBIN, T.R. 1992. Temporal coding of pheromone pulses and trains in Manducta sexta. J. Comp. Physiol. A( 171 ):505-512.

MAUELSHAGEN, J., and GREGGERS, U. 1993. Experimental access to associative learning in honey bees. Apidologie 24:249-266.

MENZEL, R. 1990. Learning, memory and "cognition" in honey bees, pp. 238-290, b~ R.P. Kesner (ed.). Neurobiology & Comparative Cognition, Laurence Erlbaum Associates, Hillsdale, New Jersey.

MENZEL, R., ErBER, J., and MASUHr, T. 1974. Learning and memory in the honey bee, pp. t95- 217, in L. Barton Browne (ed.). Experimental Ana(vsis of Insect Behaviour. Springer-Verlag, Berlin.

PHAM-DELEGUE, M.H., MASSON, C., ETII~VANT, P,, and AzAr, M. 1986. Selective olfactory choices of the honey bee among sunflower aromas: a study by combined olfactory conditioning and chemical analysis. J. Chem. EcoL 12:781-793,

PHAM-DELEGUE, M.H., ETII~VANT, P., and AzAr, M. 1988. A study of sunflower volatiles involved in honeybee discrimination among genotypes and flowering stages. J. Chem. Ecol. 12:781- 793.

PHAM-DELF.GUE, M.H., BAILEZ, O., BLIGHT, M.M., MASSON, C,, PICARD-NIZoU, A.L,, and WAD~ HAMS, L,J. 1993. Behavioural discrimination ot" oilseed rape volatiles by the honey bee Apis mellifera L. Chem. Senses 18(5):483-494.

ROBACKER, D,C~, FLOTTUm, P.K.. SAMMAtARO, D., and ErlcKson, F.H. 1982. Effects of climatic and edaphic factors on soybean flowers and on the subsequent attractiveness of the plant to honey bees, Field Crop Res. 122:267-278.

THtERY, D.. BLUET, J.M.. PHAM-DELI~GUE, M.H., ETIEVANT, P., and MASSON, C. 1990. Sunflower aroma detection by honeybee: Study by coupling gas chromatography and electroantennogra- phy. J. Chem. Ecol. 16(3):701-711.

TOLLSTEN, L., and BERGSTROM, G. 1988. Headspace volatiles of whole plants and macerated plant parts of Brassica and Sinapis, Phytochemistry 27(2):4013-4018.

VArESCHh E. 1971. Duftunterschiedung bei der Honigbiene--Einzelzell-abteilungen und Verhat- tensreaktionen. Z. Vergl. Physiol. 75:143-173.

VET, L.E., LEW~S, W.J., PAPAJ, DR., and VAN LENTEREN, J.C. 1990. A variable-response model for parasitoid foraging behaviour~ J. Insect Behav, 3(4):471-490.

WADHAMS, L.J., ANGST, M.E., and BLIGHT, M.M, 1982. Responses of the olfactory receptors of Scolytus scolytus (F.) (Coleoptera: Scolydae) to the stereoisomers of 4-methyl-3-heptanol. J. Chem. Ecol. 8:477-492.

WALt,ER, G.D,. LOPEr, G.M., and BERDEL, R.L. 1974. Olfactory discrimination by honeybees of terpenes identified from volatiles of alfalfa flowers. J. Apic, Res. 13:191-197.