• Since the 1970s it has been known that the nursery pollinator Hadena bicruris is attracted to the flowers of its most important host plant, Silene latifolia, by their scent. Here we identified important compounds for attraction of this noctuid moth.
• Gas chromatographic and electroantennographic methods were used to detect compounds eliciting signals in the antennae of the moth. Electrophysiologically active compounds were tested in wind-tunnel bioassays to foraging naïve moths, and the attractivity of these compounds was compared with that to the natural scent of whole S. latifolia flowers.
• The antennae of moths detected substances of several classes. Phenylacetaldehyde elicited the strongest signals in the antennae, but lilac aldehydes were the most attractive compounds in wind-tunnel bioassays and attracted 90% of the moths tested, as did the scent of single flowers.
• Our results show that the most common and abundant floral scent compounds in S. latifolia, lilac aldehydes, attracted most of the moths tested, indicating a specific adaptation of H. bicruris to its host plant.
In general, the insects attracted pollinate the flowers, and are rewarded with food such as pollen, nectar or oil (Vogel, 1983). However, there are also examples in which the floral volatiles themselves are the reward (Eltz et al., 2003), or in which volatiles mimic female sex pheromones of hymenopteran insects to attract males, which ‘copulate’ with the flower and thereby pollinate it (pseudocopulatory pollination). In the latter type of system, the plants exploit the male pollinators because they do not provide any reward (Schiestl et al., 2000). There are also systems in which pollinators reproduce within the flowers they pollinate (nursery pollination, see review by Dufay & Anstett, 2003). One of these systems is the interaction between the noctuid moth H. bicruris and S. latifolia (Caryophyllaceae). Silene latifolia is a perennial dioecious weed, often growing along waysides. It is naturally distributed in northern Africa, Europe and Asia (Seybold, 1990). It was accidentally introduced to North America c. 200 yr ago, followed by rapid expansion throughout the whole country (McNeill, 1977).
Hadena bicruris has dramatic effects on the fitness of its host plant in Europe, and is responsible for the destruction of c. 25% of all S. latifolia fruits produced (Wolfe, 2002). Floral volatiles are known to be important cues for H. bicruris to locate the flowers of S. latifolia. Floral scent also triggers landing on the flowers (Brantjes, 1976b). However, it is not known which floral scent compounds are important in attracting of H. bicruris. The floral scent of S. latifolia is characterized by containing several fatty acid derivatives, benzenoids and monoterpenoids, the most abundant compounds being lilac aldehyde isomers, veratrole and benzyl acetate (Jürgens et al., 2002). However, a study on the variability of floral scent in S. latifolia revealed different chemotypes (Dötterl et al., 2005). Some compounds, such as lilac aldehyde isomers, phenylacetaldehyde, veratrole and benzyl acetate, were abundant in particular chemotypes, but were only minor or even absent in other chemotypes. Behavioural studies are needed to determine the role of single compounds in attracting the moths. However, identification of potential active compounds is a necessary first step in reducing the time-consuming testing of all compounds in complex scent blends. A very useful method for this purpose is gas chromatography coupled to electroantennographic detection (GC–EAD, Arn et al., 1975; Weißbecker et al., 1997). Because behaviourally active compounds generally elicit signals in GC–EAD analyses (Schiestl & Marion-Poll, 2001; Schütz, 2001), testing only the electrophysiologically active compounds is an efficient approach. This general procedure was successfully applied in previous studies (Schütz et al., 1997; Omura et al., 1999b; Omura et al., 2000b; Schiestl & Ayasse, 2000; Schiestl et al., 2000; Plepys, 2001; Schütz & Weißbecker, 2003).
In the present study, a gas chromatograph coupled in parallel to an electroantennographic detector and a flame ionization detector (GC–EAD/FID) and a gas chromatograph coupled in parallel to an electroantennographic detector and a mass spectrometric detector (GC–EAD/MS; Weißbecker et al., 2004) were used to elucidate the floral scent compounds of S. latifolia that elicit signals in the antennae of H. bicruris. In addition, synthetic standard compounds present in the floral scent of other Caryophyllaceae species (Knudsen & Tollsten, 1993; Jürgens et al., 2002, 2003) were tested. Electrophysiologically very active compounds were tested in wind-tunnel bioassays, and the attractivity of these compounds was compared with the attractivity of the scent of whole flowers of S. latifolia.
Materials and Methods
Plant material and volatile collection
Floral scent was collected from 10 different specimens of Silene latifolia Poiret ssp. alba (Miller) Greuter & Burdet and one specimen of Silene vulgaris (Moench) Garcke, and the extracts obtained were tested with H. bicruris in GC–EAD/FID and GC–EAD/MS analyses. The plant specimens originated from seeds of different populations. Plants were grown in the glasshouse for c. 8 wk until they built up a rosette, and then placed in flower beds. For each sample, floral scent was collected from 10–40 flowers for c. 3 h using dynamic headspace methods.
The living flowers were enclosed in glass cylinders and the volatiles emitted were trapped in an adsorbent tube (pasteur pipette filled with 100 mg of a 1 : 1 mixture of Tenax-TA 60–80 and Carbotrap 20–40). The air was sucked from the glass cylinder over the trap by means of a membrane pump (G12/01 EB, Rietschle Thomas, Puchheim, Germany). Samples were collected at night, when S. latifolia is emitting most of its floral volatiles (Jürgens et al., 2002; Dötterl et al., 2005). Volatiles were eluted with 200–300 µl acetone.
An H. bicruris Hufn. (Lepidoptera: Noctuidae) culture was established by collecting eggs laid in S. latifolia flowers in the surroundings of Bayreuth (Germany). A photoperiod of 18 h light, 6 h dark and a temperature of 26°C (in light) and 18°C (in darkness) yielded at least three generations per year. The larvae were reared on freshly collected fruits of S. latifolia during all five larval stages, or were fed with artificial diet similar to that described by Shorey (1963); Shorey & Halle (1965), from the beginning of the second larval stage. It was necessary to add dried fruits of S. latifolia to the basic medium, otherwise the mortality of second-instar larvae was high (e.g. 500 g white beans, 280 g dried fruits, 9 g ascorbic acid, 12 g methyl-p-hydroxy benzoate, 6 ml formalin, 130 g agar, 1.5 l water). Naïve female and male moths, 2–7 d old, were used for the experiments. Adults were fed with a sugar solution (30%, same amounts of fructose and glucose, corresponding to S. latifolia nectar as described by Witt et al., 1999).
The composition of the floral scent solvent extracts (1 µl per sample) used for the electroantennographic detections was analysed on a Varian Saturn 2000 mass spectrometer and a Varian 3800 gas chromatograph fitted with a 1079 injector (Varian Inc., Palo Alto, CA, USA). The injector split vent was opened (1 : 10) and the injector heated initially at 150°C. The injection temperature increased during the injection from 200°C min−1 to 250°C, and was held for 2 min. A ZB-5 column was used for the analyses (length 60 m, inner diameter 0.25 mm, film thickness 0.25 µm; Phenomenex, Torrance, CA, USA). Electronic flow control was used to maintain a constant helium carrier gas flow of 1.8 ml min−1. The GC oven temperature was held for 2 min at 40°C, then increased at 5°C min−1 to 240°C and held for 3 min. The MS interface was heated to 260°C and the ion trap worked at 175°C. The mass spectra were taken at 70 eV with a scanning speed of one scan per second from m/z 40–350.
To quantify the amount of each volatile in the blend, known amounts of lilac aldehydes, trans-β-ocimene, cis-3-hexen-1-yl acetate, benzaldehyde, phenylacetaldehyde and veratrole were injected, and the mean peak area of these compounds was used for quantification.
Experiments were performed by means of a GC–EAD/FID system and a GC–EAD/MS system with different floral scent extracts of S. latifolia, and with additional standard compounds. Standard compounds were chosen from the scent of S. latifolia, and also from compounds found by Jürgens et al. (2002, 2003) in the floral scent of other Caryophyllaceae species. Antennae from seven females and two males (one antenna per specimen) of H. bicruris were tested. In addition, electroantennographic (EAG) experiments were performed with dilution series of standard compounds (Schütz et al., 1999) to compare the sensitivity of H. bicruris to different compounds, and to obtain dose–response curves for the compounds tested. Antennae from two to four moths were used in this experiment for each compound to obtain mean response values.
The GC–EAD/FID system consisted of a gas chromatograph (Vega 6300-01, Carlo Erba, Rodano, Italy), an EAD interface (Syntech, Hilversum, the Netherlands) and an EAG-amplifier (Professor U.T. Koch, University of Kaiserslautern). The GC was equipped with a capillary column (length 30 m, inner diameter 0.32 mm, DB-1), an FID and a split–splitless injector. The injector was operated in the splitless mode. Nitrogen was used as carrier gas at a pressure of 100 kPa (gas flow 2.5 ml min−1, gas vector 50 cm s−1 at 50°C). The following temperature programme was employed: start at 70°C, ramp 10°C min−1, end temperature 200°C. A split connection at the end of the capillary column allowed division of the GC effluent into two capillaries leading to the FID (35 cm, 0.2 mm ID) and the EAD (45 cm, 0.32 mm ID), respectively. This provided a split ratio of 1 : 5 (FID/EAD). The EAD interface (Weißbecker et al., 1997) was under thermostat control and guided the capillary column through the wall of the GC oven, thus preventing condensation of the sample in the cooler segments of the column. The end of the column projected into a small chamber where the effluent was mixed with cold, humidified air (400 ml min−1). A polytetrafluoroethylene (PTFE)-clad flow tube led the airflow to the detector cell, where the antenna of the moth was fixed in a perspex holder modified according to Färbert et al. (1997); Schütz et al. (1999). Both ends of the antenna were contacted to Ag/AgCl electrodes via hemolymph Ringer solution (Kaissling & Thorson, 1980). EAG potentials of the antenna were amplified by a factor of 100 and recorded with a Chromstar data acquisition system (Bruker, Bremen, Germany).
The GC–EAD/MS system (Weißbecker et al., 2004) consisted of a 6890 N gas chromatograph and a 5973 N quadrupole mass spectrometer (both Agilent, Palo Alto, CA, USA). The GC was equipped with a type 7163 autosampler, a split–splitless injector, and a J&W Scientific HP-5MS column (Agilent; length 30 m, inner diameter 0.25 mm, film thickness 0.25 µm). The injector was operated in the pulsed splitless mode. Helium (purity 99.999%) was used as carrier gas at a constant flow of 1 ml min−1, gas vector 24 cm s−1. The following temperature programme was employed: start: 50°C, hold for 1.5 min, ramp 6°C min−1 to 200°C, hold for 5 min.
A GRAPHPACK 3D/2 flow splitter (Gerstel, Mülheim, Germany) was used to split the effluent from the column into two pieces of deactivated capillary leading to the mass spectrometer (length 1 m, ID 0.1 mm) and to the EAD setup (length 1 m, ID 0.2 mm). The restriction of these capillaries resulted in an equal split of the gas flow into the two setups.
The mass spectrometer used electron ionization at 70 eV, and is operated in the scan mode with a mass range from 35 to 300 mass units at a scan speed of 2.78 scans s−1.
The EAD setup used a modified version of an ‘olfactory detector port’ (Gerstel, Mülheim, Germany). This incorporated a flexible heating sleeve (230°C) which guided the capillary out of the GC oven where the effluent of the capillary was mixed with humidified air. The airflow was directed to the insect antenna, which was housed in a detector cell made of PTFE. This setup is referred to hereafter as the EAD interface.
For recording EAG potentials, the same setup as described above was used. In addition, the amplified signal passed through a high-pass filter with a cut-off frequency of 0.01 Hz to suppress the slow drift often observed in the EAD signal. Afterwards, a constant voltage of 0.5 V was added to the signal. These steps were necessary to match the amplifier output to the input signal range (0–1 V) of a 35900E A/D-converter (Agilent). The signal was recorded using the Agilent chemstation software.
Calibration of EAD and EAG analyses
Both setups described above allowed calibration of the employed antenna with standard odour dilutions using a calibration port in the flow tube of the EAD interface (Weißbecker et al., 2004). Odour standards were produced by drenching small pieces of filter paper (2 cm2) with dilutions of odour compounds in paraffin oil (Uvasol-quality, Merck/VWR, Darmstadt, Germany). The filter paper was put into a 10 ml glass syringe, where the odour accumulated in a concentration proportional to that of the substance in the solution and its vapour pressure, according to Henry's law (Schütz et al., 1997). A reproducible stimulus could be supplied by injecting a fixed volume (5 ml) of odour-loaded air onto the antenna.
To compensate for the interindividual variation in absolute responses (mV), the signals were normalized (corrected) by use of a standard stimulus of cis-3-hexen-1-ol at a 10−3 dilution in paraffin oil. The responses to cis-3-hexen-1-ol were set as 100%.
Authentic standard compounds
Authentic standard compounds eliciting strong signals in the EAD study were used as olfactory stimuli in the behavioural tests (see Table 2; Fig. 2). With the exception of the lilac aldehydes and isopentylaldoxime, compounds with the highest available purity were purchased from commercial suppliers (benzaldehyde 99%, Aldrich, St Louis, MO, USA; cis-3-hexenalacetat >98%, Aldrich; decanal 97%, Fluka, Buchs, Switzerland; guaiacol > 98%, Merck; linalool 97%, Aldrich; phenylacetaldehyde 90%, Aldrich; veratrol 99%, Aldrich). Dr Roman Kaiser (Givaudan, Switzerland) provided isopentylaldoxime.
Table 2. Authentic standard compounds and doses used for wind-tunnel biotests
2 µg min−1
200 µg min−1
S, compound found in present work or previous study (Knudsen et al., 1993; Jürgens et al., 2002) in the scent of Silene latifolia; +, substance elicited ‘orientated flight’ in moths; –, no specimen displayed orientated flight; superscript digits indicate number of individually tested naïve moths.
(2E)-6-Acetoxy-2,6-dimethyl-2,7-octadienal (2) Linalyl acetate (1, 39.2 g; 0.2 mol) and SeO2 (22.2 g; 0.2 mol) were dissolved in 100 ml dioxane/H2O (9 : 1) and stirred at 80°C for 5 h. The reaction mixture was filtered through silica gel and the solvent was removed under reduced pressure. The residue was extracted three times with ether/petrolether (1 : 1). After removing the solvent, further purification was carried out by column chromatography (H/EA 9 : 1) to obtain 2 (13.0 g; 61.9 mmol; 31% yield). 1H-NMR [CDCl3; for comparison of the nuclear magnetic resonance (NMR) data the numbering of linalyl acetate was used]: δ 9.35 (s, 1H, H-8), 6.43 (m, 1H, J = 1.2/7.3 Hz, H-6), 5.92 (dd, 1H, J = 10.9/17.0 Hz, H-2), 5.16 (dd, 1H, J = 1.1/17.0 Hz, H-1), 5.14 (dd, 1H, J = 1.1/10.9 Hz, H-1), 2.32 (m, 2H, H-5), 2.05 (m, 1H, H-4), 1.99 (s, 3H, CH3CO), 1.87 (m, 1H, H-4), 1.70 (s, 3H, 3H-9), 1.55 (s, 3H, 3H-10). 13C-NMR (CDCl3): δ 195.0 (C-8), 153.7 (C-6), 141.1 (C-2), 113.8 (C-1), 82.3 (C-3), 38.1 (C-4), 23.8 (C-10), 23.5 (C-5), 9.1 (C-9).
(2E)-3,7-Dimethyl-1,6-octadiene-3,8-diol (3) To compound 2 (500 mg; 2.38 mmol) dissolved in 30 ml CH2Cl2, 40 ml diisobutylaluminum hydride (1.0 m in CH2Cl2) was added dropwise at 0°C. After 15 min the solution was poured slowly into a mixture of 100 ml concentrated HCl and 250 g ice. The organic layer was separated and the aqueous layer extracted twice with ethyl acetate. The combined organic layers were washed with saturated KHCO3 solution and saturated NaCl solution, dried with Na2SO4 and filtered through silica gel. After removing the solvent, further purification was carried out by column chromatography (H/EA 1 : 1) to obtain 3 (372 mg; 2.19 mmol; 92% yield). 1H-NMR (CDCl3): δ 5.89 (dd, 1H, J = 10.6/16.9 Hz, H-2), 5.39 (m, 1H, J = 1.2/7.4 Hz, H-6), 5.20 (dd, 1H, J = 1.1/16.9 Hz, H-1), 5.05 (dd, 1H, J = 1.1/10.6 Hz, H-1), 3.95 (s, 2H, 2H-8), 1.64 (s, 3H, 3H-9), 1.26 (s, 3H, 3H-10).
(2E)-6-Hydroxy-2,6-dimethyl-2,7-octadienal (4) To a stirred mixture of pyridinium chlorochromate (PCC) (15.0 g; 69.6 mmol) and CaCO3 (10.0 g; 100 mmol) in 250 ml CH2Cl2, compound 3 (5.6 g; 32.9 mmol) in 50 ml CH2Cl2 was added dropwise and stirred for 30 min at 0°C. The reaction mixture was filtered through silica gel and the filtrate was washed with 300 ml H2O. The organic phase was separated and the water phase washed twice with 50 ml CH2Cl2. The combined organic layers were washed with saturated NH4Cl-solution and saturated NaCl-solution, dried with Na2SO4 and filtered through silica gel. After removing the solvent, further purification was carried out by column chromatography (H/EA 1 : 1) to obtain 4 (2.86 g; 17.0 mmol; 52% yield). 1H-NMR (CDCl3): δ 9.33 (s, 1H, H-8), 6.46 (m, 1H, J = 1.4/7.5 Hz, H-6), 5.88 (dd, 1H, J = 10.7/17.2 Hz, H-2), 5.20 (dd, 1H, J = 1.1/17.2 Hz, H-1), 5.07 (dd, 1H, J = 1.1/10.7 Hz, H-1), 1.69 (s, 3H, 3H-9), 1.29 (s, 3H, 3H-10).
Lilac aldehyde (5) To a mixture of 25 ml MeOH and five drops of Et3N, compound 4 (5.66 g; 33.66 mmol) was added and stirred at 50°C for 24 h. After removing the solvent, the residue was adsorbed on silica gel and further purified by column chromatography (H/EA 12 : 1) to obtain 5 (3.8 g; 22.75 mmol; 68% yield). 1H-NMR (CDCl3): δ 9.79 (d, 1H, J = 1.2 Hz, H-8), 5.80 (dd, 1H, J = 10.7/17.2 Hz, H-2), 5.13 (dd, 1H, J = 1.2/17.2 Hz, H-1), 4.96 (dd, 1H, J = 1.2/10.7 Hz, H-1), 1.26 (s, 3H, 3H-10), 1.08 (d, 3H, J = 7 Hz, 3H-9). MS m/z (%): 168 (3, M+), 153 (26), 111 (95), 110 (46), 93 (100), 39 (33).
A 160 × 75 × 75 cm wind tunnel, similar to that described by Rojas (1999b), was used for behavioural tests. A Fischbach speed-controller fan (D340/E1, FDR32, Neunkirchen, Germany) pushed air through the tunnel with an air velocity of 0.35 m s−1. Four charcoal filters (145 × 457 mm, carbon thickness 16 mm, Camfil Farr, Reinfeld, Germany) cleaned the incoming air. The experiments were performed at night with red light illumination (< 0.01 µE) 1–3 h after the start of the dark period. The temperature was adjusted to 22–24°C.
Two different doses of the olfactory stimuli were used for the experiments. The higher dose was obtained by applying 10 µl pure compounds onto a 1 cm2 filter paper. The release rate of the compound was determined by GC–MS to be 200 µg min−1. This dose is comparable to the respective share of flower scent released by a small population of S. latifolia (200–1000 flowers). The lower dose of 2 µg min−1 was obtained by diluting the compound 1 : 100 in paraffin (Uvasol) and applying 10 µl of the paraffin dilution (Schütz et al., 1999) on filter paper. This dose is comparable to the scent emitted by two to 10 strongly scenting flowers of S. latifolia (Dötterl et al., 2005). Filter paper alone (high dose), or filter paper with 10 µl paraffin (low dose), was used as control. The olfactory stimuli were offered behind polyester gauze and different metal grids, so that the moths could not see the filter paper.
The attractivity of single compounds was evaluated through comparison with that of single flowers of S. latifolia. A total of 147 naïve moths (70 males, 77 females) were tested on the flowers individually. Two single flowers of S. latifolia were offered simultaneously to the moths in this experiment. The flowers were placed in the upwind end of the tunnel, behind gauze and different metal grids, so that they were invisible to the moths.
Moth responses during each test in the wind tunnel were classified into three categories: (1) orientation flight, when the moth flew toward the scent source up to within 5–10 cm of the gauze at the upwind end; (2) landing, when the moth landed on the gauze (but only if this landing was preceded by orientation flight); and (3) unrolling the proboscis, if this act was preceded by landing.
A χ2 test was used in the statistica package (Version 7, 2004; StatSoft Inc., Tulsa, OK, USA) to compare the attractivity of flowers and of single standard scent compounds, and to compare the attractivity of different doses used in the behavioural responses. To compare the dose–response curve of different compounds, the antennal response was logarithmized to linearize the curves. ancova was subsequently used in statistica to test for parallelism of regression lines.
In the electroantennographic study, 58 floral scent compounds were tested on seven female and two male moths of H. bicruris. Disregarding the sex of the moths, their antennae detected at least 19 different compounds, among them eight different benzenoids, six monoterpenoids, four fatty acid derivatives and one nitrogen-bearing compound (Table 1). Some of the compounds coeluted, or eluted very close to each other, and it was not possible to differentiate electrophysiologically between these compounds (e.g. different lilac aldehyde isomers). The antennae of the moths did not respond significantly to any phenylpropanoid or sesquiterpenoid, or to the cyclic pinenes and cineoles tested.
Table 1. Delivered amount and electrophysiological detection of volatiles tested in 44 GC–EAD (gas chromatography coupled to electroantennographic detection) runs to Hadena bicruris
Number of EAD runs
Amount delivered (ng)
Volatiles arranged according to their elution on a ZB-5 column; tr, < 0.005 ng; compounds or EADs with the same superscript letter coeluted or resulted in one EAD signal, respectively; *, identification based on injection of authentic samples; Ca, compounds found in other Caryophyllaceae species (Knudsen & Tollsten, 1993; Jürgens et al., 2002, 2003) but not in Silene latifolia; ST, sesquiterpene; +, clear EAD signal; –, no EAD signal; ?, electrophysiological detection possible but verification needed.
The strongest signals in the antennae were elicited by phenylacetaldehyde (Fig. 2). In general, the EAG responses increased with increasing dose of tested compounds (dose–response curve), and there were significant differences between compounds (ancova; F(6,131) = 6.97, P < 0.001). In the special case of lilac aldehydes, antennal response increased almost log-linearly by inreasing dose, indicating that antennae detected differences in concentration in low as well as in high amounts of these compounds. This was not the case for other compounds, such as cis-3-hexen-1-yl acetate, where antennal response increased exponentially by increasing dose in the log-linear plot, such that moths detected differences in dose only when these were in high amounts.
Ten of the 19 electrophysiologically active compounds were tested in wind-tunnel bioassays, some with two different doses (Table 2). Seven out of 10 compounds attracted (orientation flight) moths (females and males), while benzaldehyde, cis-3-hexen-1-yl acetate and p-cymene did not elicit any reaction in H. bicruris.
The most attractive compounds concerning orientation flight were lilac aldehyde isomers, although they were tested only in the small dose (Fig. 3a). These monoterpenoids attracted 90% of the moths tested, similarly to the scent of single flowers. All other behaviourally active compounds attracted significantly fewer individuals (11–63%) compared with the scent of single flowers. Linalool, which was tested in the high dose, attracted especially few specimens. Guaiacol, phenylacetaldehyde and veratrole were tested in both doses, and in all three cases more moths were attracted by the higher dose. These differences were significant in the case of veratrole (χ2 = 13.7, P < 0.01) and phenylacetaldehyde (χ2 = 6.4, P = 0.01), but not in the case of guaiacol (χ2 = 1.12, P = 0.29).
In terms of landing response, the lilac aldehydes elicited more landings on the upwind end of the wind tunnel than any other compound tested, with 65% moths responding. Moreover, they elicited even more landings than the scent of single flowers, although this difference is not significant (Fig. 3b). All other compounds elicited fewer landings than did the scent of single flowers.
Phenylacetaldehyde in the high dose was the most active compound concerning extension of the proboscis (Fig. 3c). Of the 50 moths tested, 22% were attracted by this benzenoid, landed on the upwind end of the tunnel and extended the proboscis. In contrast to the behavioural responses orientation flight and landing, proboscis extension was not elicited by all compounds tested. Only 5% of the moths tested unrolled the proboscis when offered the scent of single flowers.
In the present work, the noctuid moth H. bicruris responded strongly to some floral scent compounds of its host plant, S. latifolia, in both the electrophysiological study and the behavioural tests. However, some compounds did not elicit behavioural responses although they caused strong responses in the antennae of the moths. Altogether, this study revealed an expected electrophysiological and behavioural adaption of H. bicruris to its most important host plant.
The most attractive compounds in the behavioural tests were the lilac aldehyde isomers, which were as attractive as the scent of single flowers of S. latifolia, and to which the moth antennae were very sensitive, with changes in the dose being detected even when the compounds were in very small concentrations. Moreover, these oxygenated monoterpenes behaviourally attracted some specimens when offered at a very low dose of c. 0.2 ng min−1 (data not shown), which suggests that these compounds are very important for attracting H. bicruris to flowers of S. latifolia from some distance. These compounds are typically found in high amounts in only a few nocturnal plant species (Knudsen et al., 1993; Jürgens et al., 2002), and in low amounts also in butterfly-pollinated plants (Andersson et al., 2002). They were also the most attractive compounds for another noctuid species, Autographa gamma (Plepys et al., 2002), and are known to elicit responses in the antennae of different butterfly species (Andersson, 2003b). It is probable that these compounds are generally very important attractants for lepidopteran species, though more behavioural studies are needed. Lilac aldehydes are oxygenated monoterpenes with three chirality centres, therefore eight different isomers are possible. They were synthesized (Fig. 1) starting from an isomeric mixture of linalyl acetate, resulting in both enantiomeric isomers of lilac aldehydes A–D. Burkhardt & Mosandl (2003), Kreck & Mosandl (2003), and Kreck et al. (2003) studied the lilac aldehydes in detail, and found only four of the eight possible optical isomers in the nocturnally pollinated Syringa vulgaris. Nothing is known about the exact isomers occurring in the floral scent of S. latifolia. The optical pure isomers were not available, and further research is needed to clarify the role of single isomers in the attraction of H. bicruris.
Veratrole and phenylacetaldehyde also attracted H. bicruris, and there were considerable differences between the attractivity of different doses. High doses of these compounds (comparable with the scent emitted from a small population of S. latifolia) attracted significantly more moths than low doses (comparable with the scent of single flowers), which emphasizes the importance of compound concentrations in the attraction of moths such as H. bicruris (see also Brantjes, 1976a). These compounds may be of importance to both long-range attraction of H. bicruris to a population of flowering plants, and short-range orientation in the direct vicinity of S. latifolia flowers; in both cases the concentration of scent compounds is high. For A. gamma, Plepys et al. (2002) found concentration of compounds to be an important attraction factor, whereby doses higher or lower than the ideal dose inhibited attraction. Similar conclusions were drawn by Schütz (2001) for the attraction of the Colorado potato beetle Leptinotarsa decemlineata to potato plants. Relatively little is known about the attractivity of veratrole, which is found in few nocturnal plant species (Knudsen et al., 1993), in other insects. It elicits antennal responses and is discussed as an aggregation pheromone in locusts (Niassy et al., 1999); is known to be attractive to the southern corn rootworm Diabrotica undecimpunctata howardi (Lampman et al., 1987); and is further described as a compound eliciting feeding response in the oriental fruit fly Dacus dorsalis (Metcalf et al., 1975). It is particularly interesting that phenylacetaldehyde in high doses elicited significantly more proboscis extensions in H. bicruris than did the scent of single flowers. Given that proboscis extension can be interpreted as nectar-drinking/searching behaviour, phenylacetaldehyde might be a key compound in orientation of H. bicruris when the moths are close to or at the flowers, where the concentration is generally highest. Phenylacetaldehyde is a widespread floral scent compound (Knudsen et al., 1993), and is known as a very attractive compound for butterflies and various moth species (Cantelo & Jacobson, 1979; Haynes et al., 1991; Heath et al., 1992; Honda et al., 1998; Omura et al., 1999a, 1999b; Meagher, 2001, 2002; Cunningham et al., 2004). It is also a potential male sex pheromone of different lepidopteran species (Bestmann et al., 1977; Honda, 1980).
Another attractive compound was isopentylaldoxime, which is typically found in moth-pollinated species (Knudsen et al., 1993; Kaiser, 1994). To our knowledge, this nitrogen-bearing volatile has not yet been tested alone in behavioural bioassays.
Not all compounds tested in the wind tunnel attracted moths, although they elicited significant signals in the antennae of H. bicruris (Tables 1 and 2; Fig. 2). Two of these compounds are benzaldehyde and cis-3-hexen-1-yl acetate. Benzaldehyde is typically a minor compound in the flower scent of S. latifolia and is abundant only in exceptional cases (Dötterl et al., 2005). This benzenoid is a very widespread plant-derived compound, often found in floral volatiles (Knudsen et al., 1993). Because benzaldehyde is not very specific for S. latifolia, its lack of attraction for H. bicruris seems reasonable, given that the probability of finding S. latifolia when relying on benzaldehyde may be very low. Similarly ineffective in attracting moths is the fatty acid derivative cis-3-hexen-1-yl acetate, which is a common green leaf volatile (Pare & Tumlinson, 1999) and not a typical floral scent compound of S. latifolia (Dötterl & Jürgens, 2005). Therefore it is not surprising that this compound does not attract moths in search for flowers.
As reported by Dötterl et al. (2005), the floral scent of S. latifolia was highly variable and different chemotypes have been documented, the most frequent of which was dominated by lilac aldehyde isomers. Interestingly, these compounds are also the most critical for attraction of the moths, suggesting that H. bicruris has a chemical-based attraction that is well adapted to its most important host plant, and it appears to be advantageous to rely on these marker compounds, not on a particular scent profile. Moreover, lilac aldehydes are not often found in floral scents (Knudsen et al., 1993), increasing the probability that H. bicruris will find S. latifolia when using the lilac compounds as key volatiles for host-plant location, together with other typical scent compounds of this species such as phenylacetaldehyde, veratrole or isopentylaldoxime.
The authors thank Sigrid Liede-Schumann for supporting this study. Konrad Fiedler gave essential advice on statistical analyses, and Taina Witt gave valuable comments on the manuscript. Jette Knudsen and Roman Kaiser made authentic standards available and helped to identify unknown compounds. The comments of two anonymous referees were valuable. S.D. was supported by the German Research Foundation (Research Training Group 678).