Fig. 5.2. Bronstein's (1995) pollinator landscapes, with distributions of foraging classes and representative systems for each interaction. References:1 Paul us & Gack (1990);2 Pellmyr etal. (1996);3 Wiebes (1979);4 Young (1986); 5 Schemske & Horvitz (1984); 6 Waser (1982); 7 Herrera (1987);8 Murawski & Gilbert (1986);9 Waser (1979);10 Fleming etal. (1996).
(Janzen 1971; Ackerman et al. 1982), despite the fact that male euglossines' attraction to fragrance is an axiom of Neotropical biology (Williams & Whitten 1983). Odor-guided distance attraction of hawkmoths and beetles between isolated rainforest plants has been likened to traplines (Schatz 1990), and features similar pollen dispersal distances (Young 1986; Nilsson & Rabakonandrianina 1988), but there is no evidence that individual hawkmoths or beetles repeat daily foraging routes or learn landmarks.
Density-dependent visual foragers are attracted from close range to aggregated floral displays (Rathcke 1983) and may use fragrance as a distance attractant, a feeding cue, or a conditioning stimulus. Flowers pollinated by such foragers produce combinations ofolfactory and visual cues that may also attract herbivores or thieves (Dafni 1984). Because nectar and fragrance production incur metabolic costs (Pyke 1991) and attract unwanted plant enemies (Baldwin et al. 1997), autogamous species with reduced flowers should lose nectar and/or scent, while nectarless or scentless deceptive flowers are predicted to evolve under several scenarios (Knudsen & Tollsten 1993; Armstrong 1997). Depending upon its role in pollinator attraction, post-pollination fragrance emission might be (1) maintained to contribute to distance attraction (Eisikowitch & Rotem 1987), (2) eliminated to prevent futile visits (Tollsten 1993), or (3) modified to promote learned avoidance of reward-depleted flowers (Lex 1954). The chemical diversity, perceptual complexity, and salience of fragrance blends as learned cues suggest that species-specific fragrances should promote floral constancy, but direct tests of this hypothesis are needed.
Insects that visit flowers for some element of their own sexual reproduction include obligate mutualists and the victims of sexual- and brood site-deception (Faegri & van der Pijl 1979). Here, fragrances may elicit hard-wired sexual- or oviposition-related behaviors, although feeding may also be involved (Beaman et al. 1988). Interactions between figs and fig wasps are governed by scent from a distance and by contact chemore-ception after landing (Gibernau & Hossaert-McKey 1998). Initial data support the hypothesis of ethological isolation through species-specific fragrances (Ware et al. 1993; Grison et al. 1999). In all documented cases of pseudocopulatory pollination, fragrances effectively mimic female hyme-nopteran sex pheromones (Borg-Karlson 1990). Peakall (1990) proposed that pollinator movement among sexually deceptive orchids should reflect optimal mate search, not optimal food foraging. Male thynnine wasps that pollinate Australian Caladenia and Cryptostylis orchids show site aversion after an attempted mating, promoting greater dispersal distances and fewer local visits than would be the case during nectar foraging (Peakall & Beattie 1996). In contrast, the post-copulatory avoidance of females and Ophrys flowers by male andrenid bees is based on aversive odor learning (Ayasse et al. 1993), resulting in bee movement toward novel odors and subtle fragrance variation among adjacent flowers (Schiestl et al. 1997).
Pollinator-mediated selection on variation in floral scent
This chapter's emphasis on pollinator attraction reflects an historical adaptive bias in fragrance research, but recent studies have explored alternative roles for floral volatiles. Pellmyr & Thien (1986), Armbruster (1997), and Schiestl et al. (1999) argue that plant defense and stress-response physiology are pre-adaptive for fragrance evolution. Herbivory on crop plants induces systemic vegetative emissions that are chemically "floral" and that attract parasitoids (Pare & Tumlinson 1997). Less appreciated are the deterrent components of fragrances, whether they combat microbes in standing nectar (Lawton et al. 1993) or repel unbidden floral visitors (Omura et al. 2000).
The potential for positive directional selection (by pollinators) on fragrance variation is supported by the following evidence. Galen & Kevan (1983) characterized "sweet" and "skunky" fragrance morphs of Polemonium viscosum, showing that bumble bees overvisit sweet scents and discriminate against skunky scents, irrespective of nectar quantity. Furthermore, sweet fragrance was correlated with floral geometry and with nectar traits favorable to bumble bee pollination (Galen & Newport 1987). Pellmyr (1986) documented scent-based assortative visitation by butterflies and bumble bees to specific Cimicifuga simplex chemotypes; he proposed that incipient speciation could occur through floral isolation. Finally, Pelz et al. (1997) discovered that increased odor concentration promotes more salient conditioning and memory consolidation in bees, at least for single compounds.
Negative directional selection (by florivores) on fragrance is supported by equally strong evidence. Galen (1983, 1999) demonstrated that ants destructively overvisit sweet-scented Polemonium flowers, stealing nectar and detaching styles. Ecroyd (1996) investigated the role of scented nectar in bat pollination of Dactylanthus taylorii in New Zealand and found that introduced opossums and rats were attracted by the fragrance and
Was this article helpful?