Naturalists have long wondered why the shape and colour of the flowers of Ophrys are similar to those of certain insects4. The odour of a flower was found to be the most important signal triggering pseudocopulation behaviour2. Although many odour components were produced by the flowers as well as by their pollinators5,6,7, there has been no convincing identification of the compounds that trigger pseudocopulation. We have used gas chromatography-electroantennographic detection to identify physiologically active compounds, which were subsequently tested in the field for behavioural activity. The activity of synthetic mixtures was compared with that of the sex pheromone of female bees and the pollinator-attracting scent of Ophrys flowers.

We tested various odour samples of female bees and orchid flowers for their attractiveness to male bees (Fig. 1). Of all bee-odour samples tested, cuticle extracts elicited the most intense behavioural reactions from males, indicating that this is where the sex pheromone of A. nigroaenea is located. Extracts from the Ophrys flower labellum also elicited frequent copulation attempts by males (Fig. 1).

Figure 1: Reactions of Andrena nigroaenea males in behavioural tests in the field.
figure 1

Odourless dummies (Soxhlet-extracted, dried A. nigroaenea females) scented with different samples were offered to A. nigroaenea males in a patrolling area for 3 min. Means (±s.e.m.) of approaches to the dummy to less than 5 cm (pale blue bars), pouncing on the dummy (royal blue bars), alighting on the dummy (beige bars) and copulation attempts with the dummy (dark blue bars) are plotted. The Mann-Whitney U -test was used for statistical analysis. Asterisks indicate significant differences between the reaction types of the odourless-dummy group and each of the other six test groups at significance levels of *P <0.05 and **P <0.001.

In the experiments using gas chromatography-electroantennographic detection, 15 components from the attractive odour samples of female bees triggered electro-antennographic responses in male antennae (Table 1). The Ophrys flower extracts contained all except one of these compounds. The compounds were identified by gas chromatographic-mass spectrometric analy-ses as straight-chain saturated and unsaturated hydrocarbons with 21-29 carbon atoms. Both sample types contained these compounds in very similar relative proportions (Table 1). To investigate the behavioural activity of electrophysiologically active compounds, we performed behavioural tests with male bees. A blend of 14 synthetic hydrocarbons occurring in both the bee scent and the flower scent, mixed in accordance with the relative proportions found in either sample type, elicited frequent copulation attempts in males when applied to a dummy (Fig. 1).

Table 1 Electrophysiologically active compounds in cuticle extracts of virgin Andrena nigroaenea females and labellum extracts of Ophrys sphegodes flowers

Our understanding of the mechanisms of chemical mimicry in Ophrys orchids has been constrained by a limited knowledge of the chemical composition of sex pheromones in the solitary bees and wasps that serve as pollinators. Earlier studies reported aliphatic 1- and 2-alcohols and terpenes in flowers and bee odours to be the key components for pollinator attraction in several Ophrys species5,7. However, these compounds did not elicit definite copulation attempts in male bees. This led to the hypothesis that Ophrys flowers produce only a set of “second-class attractivity compounds” that attract only the part of the pollinator population with a low threshold for sexual stimuli5.

Our results challenge this view. In O. sphegodes, we find that the behaviour-inducing odour bouquet is more or less identical to the sex pheromone of the female bees and therefore elicits nearly equally intense reactions. But the behaviourally active compounds are ubiquitous or occur in at least trace amounts in plant cuticular waxes8. The similarities in their relative amounts in orchid labella and attractive female bees supports the view that the specific pattern of otherwise common compounds is the key innovation of O. sphegodes for the attraction and deception of the pollinators.

These findings shed new light on the evolution of this fascinating insect-plant relationship. Cuticular hydrocarbons, as a part of the plant surface wax, primarily prevent the loss of water. During the evolution of chemical mimicry in O. sphegodes, these compounds obtained the additional function of attracting pollinators, which was achieved through alterations in their relative proportions. It seems feasible that, in an ancestor of Ophrys, the pattern in a mutant occasionally resembled that of the sex pheromone of a pollinator species. This could have led to pollination by sexually excited males of this species, and natural selection would have favoured further plant mutants with a hydrocarbon pattern with an even closer resemblance. Apart from advantages discussed elsewhere9, being pollinated by pseudocopulating males allowed the plant to decrease the costly emission10 of typical floral-odour compounds. Flowers of O. sphegodes, like those of many Ophrys species, emit only minute quantities of such volatiles5,11. Chemical mimicry in the sexually deceptive pollination system therefore also provided an economical way for the plant to ensure its transfer of gametes.