Abstract

Pollination by insects is largely unique to the angiosperms, and is presumably responsible for the abundant success of this group of plants. Floral morphology, color, and pollinator rewards have often been characterized in their roles of adaptations to different groups of pollinators. Floral scents and odors, however recognized as valuable chemotaxic markers for pollinators, are less well researched. While known to play major roles as secondary attractants for pollinators, those that are well studied are arbitrarily considered to produce heady and pleasant aromas and which serve as chemical attractants for pollinators within the Hymenoptera and Lepidoptera. Recent and substantial research, however, is committed to the pollination ecology of those plant species that emit unpleasant odors (frequently equated to the odors of rotting fish, carrion, or dung), which serve as important chemical attractants for insects within the orders Coleoptera and Diptera. The presence of these odors is generally assumed to be due to the assemblage of amine-containing compounds. Analysis of these odors through headspace adsorption and solvent extraction for sample preparation followed by capillary GC and GC-MS analysis further indicates the presence of indole, p-cresol and 2-heptanone which impart fecal and urinous characters. Several biological and ecological functions are important in regulating the mode and timing of release of these odors to their pollinators. Among these are thermogenecity, diurnal variation in scent production, correspondent release with daily and seasonal insect activity, and temperature. From an evolutionary perspective, it is suggested that pollination in early angiosperms often was based on the intertwining of the sexual life cycles of insects with plants in which floral odors served as chemical cues for mating sites (the flower) and food (usually pollen). Finally, the ecological significance and advantage of pollination by species in the Diptera and Coleoptera through broadcast of these odors is reviewed.

Introduction

That plant scents and odors have been recognized as valuable chemotaxonomic markers is well established (Harborne, 1993). However, the majority of floral odor research has been focused on the rather pleasant and heady aromas that serve as chemical attractants for insect species within the Lepidoptera and Hymenoptera. For pollination ecology studies, the focus is often centered on the role that flower morphology and flower color play in pollination ecology with scent recognized as a secondary player.

Floral scent is an important component in the reproductive biology of many flowering plants (Harborne, 1993). Scents will serve as attractants in plants pollinated by a variety of fauna: birds, bats, beetles, butterflies, moths, bees and wasps. Scent will advertise the presence of awards to foraging pollinators, such as nectar or pollen, or temporary protection from predators. Research in floral scent ecology has important implications for studies in population ecology, pollen dispersal, pollination ecology, plant speciation, insect behavior, and pest management (Raguso and Pellmyr, 1998). Further, Pellmyr and Thien (Pellmyr and Thien, 1986) suggest that pollination systems between insects and extant archaic angiosperms evolved primarily through the meshing of the sexual life cycles of phytophagous insects with flowers.

In light of the recognition that floral scent plays a key role in pollination ecology, much recent research has been focused on the pollination ecology of those plant species that emit arbitrarily unpleasant odors, often likened with the odors of decaying tissues and fecal material. The composition of these odors serves as important ecological cues for insects within the Coleoptera and Diptera. Beetle-pollinated flowers, for example, often provide the pollinators with a space to protect them from predators, and/or produce relatively large amounts of pollen and sometimes special nutritive tissues to provide nutrition for the pollinators (Beach 1982, Gottsberger 1990). In addition to cueing for the presence of pollinator awards (nectar, pollen) and protection, these odor compounds will also often cue for mating and oviposition sites which can result in an exercise of deceit pollination where the pollinator leaves the plant often without benefit of reward. Additionally, these pollination systems may have evolved for plants which inhabit areas which are unsuitable for hymenopteran and lepidopteran species, such as the littoral layers of forests or in rocky and treeless areas with few flowering plant species.

Odoriferous chemical compounds in plants

Foul and fetid odors produced by plants are generally associated with a pollination syndrome known as sapromyophily, or pollination by insects that breed or feed on decaying matter or fungi (Kite and Smith, 1997). In most beetle-and some fly- pollination schemes, scent acts as the primary attractant (Gottsberger 1990). Recent work on odoriferous compounds suggest that the chemical cues emitted from Sauromatum guttatum (Araceae), for example, may be more important to these insects than the visual cues (Skubatz et al., 1996). More precisely, the physiological features of the plant that actually produce the scent (described below) are capable of dispersing the scent over long distances.

The Sauromatum appendix was found to emit a complex blend of terpenes, fatty acid derivatives, phenolic, sulphur and nitrogenous compounds (Skubatz et aL., 1996). Indole and 30 other nitrogenous compounds were identified, to which over 30 insect species, many of which were species of Coleoptera and Diptera, were attracted (Skubatz et aL., 1996). Among these were compounds from at least nine different chemical classes. The major group in those compounds released were terpenes (mono- and sesquiterpenes), with monoterpenes accounting for approximately 70% of the mixture, and sesquiterpenes accounting for about 6% of the mixture. As a group, fatty acids, alcohols, ketones and aldehydes accounted for about 20% of the mixture. Those compounds responsible for the characteristic fetid odor of the plant, indole and sulphur compounds were found in 0.2% and trace amounts, respectively (Skubatz et aL., 1996). The nitrogen and the sulphur-containing compounds provide the distinctive vile aroma of the appendix. For example, indole is pungent and dimethyl disulphide is intensely foul (Skubatz et aL., 1996).

The foul odor produced by these plants is apparently attractive to many species of Coleoptera and Diptera that are associated with the ecological recycling of dung and carrion (Skubatz et aL., 1996). The affinity these insects have for these compounds released by the plant is not surprising since the same volatiles which are released from Sauromatum guttatum are also the same as those liberated from decomposing tissue and wastes through microbial oxidation of lipids and decarboxylation of amino acids. Interestingly, terpenes, which are the prevalent pleasant-smelling compounds in the majority of flowering species, are also found in the Sauromatum guttatum appendix. However, their presence is masked (to the human nose) by the presence of the malodorous compounds which are detectable at very low concentrations due in part to their volatility and compact molecular structures (Skubatz et aL., 1996).

Another foul-smelling plant, Arum maculatum (Araceae) possesses an odor described as foul and urinous which is produced by the presence of indole and additionally p-cresol. The combination of these two compounds in Arum maculatum mimics the pollinator’s food and oviposition sites resulting in deception (Kite and Smith, 1997). The primary pollinator for Arum maculatum is a dipteran Psychoda phalaenoids (Psychodidae) which in field observations breeds exclusively in cow dung (Kite, 1995).

Analysis of cow dung odor to compare with the pungent components of Arum maculatum revealed that among the high amounts of hydrocarbons present, p-cresol, nonanal and decanal were consistently present, with p-cresol the most abundant volatile produced. Previous more detailed studies additionally detected indole, phenol, benzaldehyde, which are important volatile components of fresh dung (Kite, 1995), as well as various ketones, acids, sulphides and skatole. P-cresol is also known to be a semiochemical for several insects, and is specifically implicated as an olfactory stimulant for the Japanese dung beetle Geotrupes auratus, as well as serving as an oviposition stimulant for mosquitoes of the genus Toxorhynchites (Kite, 1995).

Deceit pollination has also been observed in Orchidantha inouei (Lowiaceae: Zingerbales) was recently discovered in Malaysia (Sakai and Inoue, 1999). This pollination system is believed to be an oddity in this particular group of plants, as many members of this family exhibit fascinating mutual relationships with their pollinators. Orchidantha inouei attracts two genera of dung beetle (Coleoptera: Scarabaeidae), only one of which was observed to actually pollinate this orchid, although the orchid itself does not offer any nutritional reward nor offer protection, hence the observation that this is deceit pollination (Sakai and Inoue, 1999).

Analysis of odors

Non-invasive headspace techniques have recently been developed to characterize floral scents and odors (Kite, 1995). The advantage to using non-invasive techniques is that the lack of disturbance to the floral anatomy is less likely to involve changes in chemical composition of the floral odors due to injury of the inflorescence. Additionally, these techniques make the production of artefacts produced by distillation or extraction less likely (Kite, 1995). These techniques involve collecting volatiles on a sorbent trap which is analyzed through gas chromatographic analysis (Kite, 1995).

The chemical composition of species of Arum previously was analyzed through invasive techniques that involved cutting of the appendices (Kite, 1995). Distillation of isobutylamine from the cut appendices under alkaline conditions, and condensation of vapors, also from cut appendices, detected several monoamines, diamines compounded with skatole. These simple amines have been generally assumed to produce the foul and fecal odors often associated with members of this genus. However, these techniques failed to specifically detect indole in Arum maculatum, although it was detected and implicated in the foul odors of other species of Araceae.

In one particular study by Kite (1995) on the floral odor analysis of Arum maculatum, the isolated volatiles using these new techniques were compared with volatiles produced by cow dung to identify any common components between the two. The analyses revealed 56 compounds in the odor consisting of 11 aliphatic aldehydes, ketones, esters or alcohols, three benzenoids, 37 terpenoids, indole (a nitrogen-containing compound) and four hydrocarbons. From these analyses, it was found that the foul odor-producing constituents of Arum maculatum were mostly due to the presence of indole, p-cresol, and 2-heptanone. This is unusual among flower scents in that these components usually comprise a very minor portion of the odors produced. In Arum maculatum, however, these sesquiterpenoids comprise a large proportion of the volatiles, about 40-50% (Kite, 1995). The pungent aroma of these three components is additionally found to be amended by the presence of spicy-scented terpenoids. These compounds are believed to impart a more alluring quality to the characteristic aroma of Arum maculatum, which may function in conjunction with the pungency to lure the pollinators into the base of the floral spathe and into the chamber, where contact with pollen is the greatest.

Other odor analyses conducted on the inflorescence of Hydrosome rivieri (Araceae) through solvent capture and gas chromatography-mass spectometry techniques identified six odor components. Among these were dimethyl disulphide and dimethyl trisulphide, which are the components with a rotting meat comparable with that of carrion (Stránský and Valterová, 1999).

Further odor analyses of the odor-producing components of Senecio articulatum (Asteraceae), a cultivated plant in South America with inconspicous flowers, indicates the presence of 3-methylbutanoic acid as a major component, accounting for 39%of the trapped volatiles (Kite and Smith, 1997). The 3-methylbutanoic acid is an uncommon constituent of floral odors, and is reported in only four other genera of plants (Kite and Smith, 1997). The odor produced by Senecio articulatus is clearly different from that of other sapromyophilous flowers in that the odor clearly resembles dung or rotting meat, that is, the odor is not amended by the presence of sweeter- smelling components (Kite and Smith, 1997). The distinction suggested here is that the production of an unattended odor indicates to pollinators that an award is imminent (pollen), whereas the odor produced by the presence of indole and p-cresol in Arum maculatum mimics the pollinators’ food and oviposition sites resulting in deception (Kite and Smith, 1997).

The role of thermogenecity in the release of odors

All known flowers contain metabolic biochemical activity, the by-product of which is heat (Seymour and Schultze-Motel, 1997). However, in most flower-producing plants, the reaction to produce heat is slow and therefore the heat dissipates slowly. The plants technically classified as "thermogenic", however, produce an unusually large amount of heat over a very short period of time. This significant heat production is assumed to not be a by-product of metabolic activity, but produced rather for the sake of completing some ecological need or function (Seymour and Schultze-Motel, 1997). This episodic heat production in these plants usually corresponds with the period when the female flower parts are most receptive to pollination and the floral scent is the strongest (Seymour and Schultze-Motel, 1997).

Thermogenic flowers are known to produce aromas that range from sweet perfumes to nauseating stenches (Seymour and Schultze-Motel, 1997). Many of the foul-odor producing species of flowers contain indoles and skatoles that mimic the pungent aroma of decaying flesh or dung. Heat generated in the appendix most certainly aids the volatilization of these higher boiling point compounds and is therefore important in influencing the chemical composition of the headspace, and hence the character of the odor (Kite, 1995).

For example, the floral odor of Arum maculatum has been frequently described as ‘reminiscent of decaying urine’ or ‘foul and urinous’ (Kite, 1995). This foul odor is produced by the appendix of the flower after the spathe has opened and rapid cyanide-resistant respiration in the appendix tissue causes the appendix itself to heat up. The thermogenic action of the flower therefore causes a volatilization of the odor which is then more easily dispersed than if no heating at all took place (Kite, 1995). Sauromatum guttatum also undergoes a heat-producing stage which aids in liberating the volatiles coinciding with the female flower’s receptivity to pollinators attracted to the appendix by the odor that is emitted ((Skubatz et aL., 1996).

Gottsberger (1990) reports that pollination occurs by large dynastid scarab beetles of the genus Cyclocephala in two plants, Annona coriacea (Annonaceae) and Philodendron selloum (Araceae). Gottsberger further states that although the two plant species are widely separated phylogenetically, their pollination biologies are very similar, which indicates that their adaptations to pollination by beetles is the result of convergent evolutionary developments.

Annona coriacea possesses a thermogenetic function in order to attract its beetle pollinator Cyclocephala atricapilla (Coleoptera: Dynastidae) (Gottsberger, 1990). Apparently this plant will produce a functional flower that warms to 34° C, which may be as much as 15° C above ambient air temperature. Again, the purpose of this thermogenesis is to volatilize a characteristic spicy odor which attracts the beetle for pollination. This takes place over the course two days. The flowers, which are protogynous, have the male and female phases of the flower distinct from one another so that self-pollination is unlikely. During the first evening, the flower enters its female phase, attracting its beetle pollinators over long distance through its thermogenetic function. The flower’s petals will close off the entrance, effectively trapping the beetles in the chamber for approximately 24 hours. The next evening, the stamens will detach from the receptacle and shed pollen grains, which cover the body of the beetle. As the petals drop, the floral chamber is opened, and the beetles are liberated. These pollen-covered beetles fly off and then enter a newly opened odoriferous flower in the female stage to successfully induce pollination. The Annona provides for its pollinator rewards such as comestibles, mating and protection sites (Gottsberger, 1990).

Philodendron selloum (Araceae) also possesses a thermogenetic function (Gottsberger, 1990). Thermogenesis occurs in the evening hours, and may reach temperatures of 46° C, which may be as much as 30° C more than ambient temperature. Philodendron selloum, however, is one of the few plant species known which has lipid oxidation during the main heating phase, insect of normal starch utilization. This very distinct respiration process, with its accompanying heat production, causes an accentuated volatilization of odor components, attracting its representative beetle pollinator, Erioscelis emarginata (Dynastidae), over long distances. The attractivity of the inflorescence is often so intense that 50, 100 or even 200 beetles may be observed approaching a single inflorescence simultaneously (Gottsberger, 1990).

In another thermogenetic example, the Amazon water lily, Victoria amazonica, heat-production is combined with a change in petal color during a two-day sequence, presumably to control the behavior of its scarab beetle pollinators of the genus Cyclocephala (Seymour and Schultze-Motel, 1997). These flowers on the first day, floresce during the evening hours, when these beetles are the most active. Cyclocephala beetles fly to the petals, crowd the floral chamber, and eat a nutritious starchy material provided by the lily. Later, the petals will close around the beetles, effectively trapping them for approximately 24 hours. The next evening, the petals reopen, allowing escape of the now pollen-covered beetles to fly to the next series of first-day flowers. The flower of escape is no longer attractive to the beetles, as the scent no longer lingers, neither the flowers are white (Seymour and Schultze-Motel, 1997).

Timing of inflorescence and pollination

Variation in the diurnal and seasonal production of floral scents is also of key importance to successful pollination of odor producing flora, and correlate well with daily and seasonal insect activity.

Amorphophallus johnsonii flowers during April in the main rainy season (Beath, 1996). Anthesis of this plant commences at dusk when emission of a strong aminoid odor takes place. Large numbers of carrion beetles of the crepuscular species Phaeochrous amplus begin to arrive just after dark and become trapped in the lower spathe overnight and remain in the appendix of the inflorescence for approximately 24 hours. At this time the anthers produce long threads of sticky pollen, which adhere to the beetles as they finally make their escape, and fly to the next aminoid-emitting inflorescence, perhaps now in its female stage, resulting in pollination of the next flower.

Experiments with Phaeochrous amplus beetles and pollination of Amorphallus johnsonii demonstrated that beetles traveled from male phase blooms to female phase blooms on the same day, and were even observed up to three days later, with distances traveled by marked beetles of eight to 37 meters (Beath, 1996). The number of successfully fertilized flowers observed in this manner was given about a 40% success rate. The reason for the success of certain flowers over others was found to be that successful fertilization only occurred when the female phase blooms were visited by beetles coming from a bloom in the male phase on the same evening (Beath, 1996).

Senecio articulatus produces an odor that is described as rather repugnant and fetid, which is more noticeable during the morning hours (Kite and Smith, 1997). Although actual pollination of Senecio articulatus has not been directly observed in the field, insects associated with they recycling of nutrients have been observed to visit the inflorescences, including calliphorid flies (Kite and Smith, 1997), which are known to be active during the daylight hours.

The timing and release of 3-methylbutanoic acid from Senecio articulatus suggests that the responsible pollinators are active during the day, which supports the assumption that calliphorid flies are presumably the pollinators (Kite and Smith, 1997).

Similarly Arum maculatum (Araceae) inflorescences usually open mid to late morning and temporarily trap small flies, the majority of which are diurnally active Psychoda phalaenoides (Diptera: Psychodidae), which are attracted to the dung-like scent emitted by the thermogenic inflorescence (Ollerton and Diaz, 1999).

Evolutionary significance and ecological functions of thermogenecity and chemoattractants

Evolutionarily speaking, the significance of thermogenesis for beetle pollination is apparent in several ways (Seymour and Schultze-Motel, 1997). First, it is apparent that thermogenic flowers are always protogynous and the spike of heat production occurs with the period of female flower receptivity. Second, the morphology of the flowers is well-suited for beetle pollination. The flowers are generally large in size, and provide broad landing platforms for clumsy fliers (beetles). Third, the floral scents often mimic those scents that beetles associate with places to feed, mate, or lay eggs, and it is believe that the fragrances have co-evolved with the beetles to induce specific activities (Seymour and Schultze-Motel, 1997). Lastly, there is evidence for correlation between the thermal requirements of beetles and the temperatures maintained inside the thermogenic flowers (Seymour and Schultze-Motel, 1997). Most beetles are endothermic and require high temperatures for activity such as flight, mate competition, and feeding. Higher and non-lethal temperatures therefore reduce the energy expenditures by beetles required by the plants for pollination. It is significant to note that the temperatures often found in thermogenic flowers are in the same range preferred by active beetles (Seymour and Schultze-Motel, 1997).

One ecological explanation for the very high temperature reached in Philodendron selloum (highest temperature measured in plants to date), for example, may be found in its population structure (Gottsberger, 1990). Apparently, Philodendron selloum inhabits semi-dry forests on the Brazilian plateau, quite in contrast with other members of the genus, which are found to inhabit humid rain forests. Therefore, suitable habitat for Philodendron selloum is scarce, and so its population density is sparse, and individual plants are often found far from one another. It seems certain that the need for such highly specialized floral biology to successfully attract pollinators over such long distances contributes to its success as a species.

Dung beetles have been observed to be excellent searchers for dung material, and many species will fly for long distances in search of a particular type of dung. Dung beetles, therefore, are suggested by Sakai and Inoue (1999) to provide long-distance pollen transfer. The interest here is that many plant species of Zingerbales specialize in long-distance yet costly (in terms of protection and nutritional rewards) pollination tactics through pollinators such as bees, birds and bats. One species within this group, Orchidantha inouei does not offer such costly rewards for pollination, and is presumed to be an energy saver (Sakai and Inoue, 1999). Orchidantha inouei (Zingerbales) attracts two genera of dung beetles, Onthophagus and Paragymnopleurus (Sakai and Inoue, 1999). However, Onthophagus, being the smaller of the two genera, was the only one observed to actually carry pollen as it left the flower, and through this particular study, is identified as the major pollinator of Orchidantha inouei. Although several individuals of Paragymnopleurus were observed to aggregate around the base of the plant, pollen on the bodies of individuals was observed rarely (two out of 27). This was possibly because Paragymnopleurus beetles are too large to reach the anthers located in the narrow, innermost part of the corolla of the flower, as was facilitated by Onthophagus’ smaller body size (observed pollen carriers 18 of 30 cases). However, the odor produced in Orchidantha inouei was not observed to attract dung or carrion flies within the dipteran families Muscidae, Calliphoridae or Scatophagidae (Sakai and Inoue, 1999). It is suggested by these observations that the odor is not a precise imitation of dung or carrion, and so do not always function to attract specific pollinators.

On the other hand, the chemical composition of foul Amorphophallus plant odors (a cocktail of amines and indole compounds volatilized by thermogenesis) do attract distinct species of pollinators. Each species of Amorphophallus studied has revealed a characteristic chemical compound and are presumed to have evolved thusly to each attract a specific pollinator (Beath, 1996). Selecting for specific pollinators increases pollen discrimination between simultaneously blooming plants and reduces the likelihood of production of wasted gametes through mixing of pollen between incompatible plants species.

Additionally, fly pollinated plants which produce a variety of odor-producing compounds can potentially exploit different aspects of a dipteran pollinator’s behavior through production of different odors conforming to various chemical types (Kite and Smith, 1997). Calliphorid flies, for example, are known to be able to perceive differences in the odor compositions of decaying meat and flowers through different antennal receptors (Kite and Smith, 1997). Therefore, production of various odoriferous compounds to exploit the perception of the calliphorids will increase the likelihood that the plant will successfully attract the fly for pollination among individuals of its species.

Literature Cited