Dependency on floral resources determines the animals' responses to floral scents

Abstract

Animal-pollinated angiosperms either depend on cross-pollination or may also reproduce after self-pollination—the former are thus obligately, the latter facultatively dependent on the service of animal-pollinators. Analogously, flower visitors either solely feed on floral resources or complement their diet with these, and are hence dependent or not on the flowers they visit. We assume that obligate flower visitors evolved abilities that enable them to effectively forage on flowers including mechanisms to bypass or tolerate floral defences such as morphological barriers and repellent/deterrent secondary metabolites. Facultative flower visitors, in contrast, are supposed to lack these adaptations and are often prevented to consume floral resources by defence mechanisms. In cases where obligate flower visitors are mutualists and facultative ones are antagonists, this dichotomy provides a solution for the plants' dilemma to attract pollinators and simultaneously repel exploiters. In a meta-analysis, we recently supported this hypothesis: obligate flower visitors are attracted to floral scents, while facultative ones are repelled. Here, we add empirical evidence to these results: bumblebees and ants, obligate and facultative flower visitors, respectively, responded as predicted by the results of the meta-analysis to synthetic floral scent compounds.

The mutualism between flowers and their pollinators is often exploited by cheaters that consume floral rewards but do not contribute to or even reduce the reproductive success of plants.¹ The classification into mutualistic and antagonistic flower visitors represents a phytocentric point of view and only considers the interaction's net effect for the plant. However, the outcome of each plant-flower visitor interaction may be highly conditional and variable over time² and thus constitutes a continuum between beneficial and detrimental, and it may not be unequivocally assigned to be either positive or negative. Furthermore, many flower visitor species that function as effective pollinators of some plant species represent severe antagonists to other plant species.³ Thus, except for highly specialised systems, it is difficult to predict whether an interaction is mutualistic, commensalistic or antagonistic. We proposed a different classification of flower visitors based on the animals' interest in flower visits.⁴ Animals visit flowers primarily in search for food; pollination is just a secondary effect.⁵ For some taxa nectar and pollen are the sole nutrient supply, others only supplement their more generalistic diets with floral resources. These different dependencies on floral resources can often be unequivocally assigned to each animal species. Bees, for example, strongly depend on pollen and nectar and are thus obligate flower visitors. In contrast, ants are omnivores and thus facultative flower visitors that consume large amounts of floral nectar of some plant species but obtain most of the nutrients required by the colony from non-floral resources.⁶

Optimal foraging theory predicts that animals evolve physiological and behavioral features that allow them to exploit their resources as effectively as possible.^7,8 Therefore, a classification considering the animals' dependencies on floral resources (obligate versus facultative) may be better suited to explain adaptations to flower visits than their effect on plants' reproduction (mutualistic versus antagonistic). One very important adaptation to the consumption of floral resources is the ability to tolerate or overcome floral defences that are employed by the flowers to reduce the visitation frequency of detrimental flower visitors.⁹ Floral scents are innate attractants or reinforce floral visits due to associative learning but do also serve as effective repellents against antagonists.¹⁰ In a meta-analysis, we recently demonstrated that the dependency on floral resources determines the responses to floral scents.⁴ In the bioassay presented here, using bumblebees (Bombus terrestris) and ants (Lasius niger), we empirically tested the predictions deduced from the metaanalysis. We expected that bumblebees—as obligate flower visitors—are attracted to floral scent compounds, while ants—as facultative flower visitors—are repelled.

Bioassay

Bumblebees (B. terrestris, one colony provided by a commercial supplier) were reared in a climate chamber (day/night: 12 h/12 h, 24°C/19°C). Ants (L. niger, two colonies studied in situ) were studied in a fallow land near the university campus in Würzburg, Germany. Workers of both species were allowed to choose between sugar-baits (15% sucrose solution) surrounded by different scents. The sugar solution was presented in 1.5 ml microcaps and dispensed by a wick that was inserted through a hole in the lid. Thirty baits were placed on a wooden board with drilled holes. Each bait was surrounded by a filter paper (∅ = 55 mm) treated with substances from 4 chemical classes (aliphates, benzenoids, mono- and sesquiterpenes) and a control. Six replicates were used per treatment. Treatments were arranged on the board in a randomized block design. Substances were solved in acetone p.A. and 200 µl of this solution containing 0.01 or 0.005 mMol of the respective substance was applied on the filter paper. Pure acetone p.A. was used as control. Three different combinations of scents were used: (1) 1-hexanol (Roth, >98%), benzaldehyde (Fluka, >99%), linalool (Merck, >97%) and nerolidol (Merck, >95%); (2) n-pentadecane (Roth, >99%), eugenol (Merck, >99%), limonene (Roth, >95%) and β-caryophyllene (Roth, >95%); (3) n-pentadecane, eugenolmethylether (Roth, >98%), citronellol (Roth, >90%) and trans-farnsesol (Aldrich, >96%). Trials with each substance combination were repeated 2–3 times per animal species (at least one time with each concentration per compound). The board with the scented baits was placed in the climate chamber and number of bumblebees visiting the baits within 1 hr was counted. For ants, the board was placed in the vicinity of a nest, after 30 minutes the filter papers were treated with scents and number of ants at baits was counted for 30 minutes in 5-minute intervals. Log response ratios $L=\ln \left({\overline{X}}_E/{\overline{X}}_C\right)$ were calculated with ${\overline{X}}_E$ = number of insects visiting each scented bait and ${\overline{X}}_C$ = mean number of animals on the six control baits per trial (L was also used in Junker and Blüthgen⁴ to which results can be compared). Positive values of L indicate attraction, negative values repellence. A three-factorial ANOVA was conducted to reveal the effects of the substances in different concentrations on the two species.

Ants negatively responded to all floral scent compounds except for aliphatics (1-hexanol and n-pentadecane). In contrast, foraging bumblebees were unaffected by individual floral scent compounds (Fig. 1). L was strongly affected by the insects and the substances tested, but not by the different concentrations used in the bioassay (Table 1).

Figure 1.

Effects of floral scent compounds on bumblebees (open circles) and ants (black squares). Data are mean and 95% confidence interval of the log response ratio L. Since different concentrations did not affect the animal's decision, they were pooled in this figure. Sample sizes are indicated above each point. A significant deviation from neutrality or other trials is indicated in cases were confidence intervals do not overlap zero or other confidence intervals. Numbers below each pair of a circle and a square are: (1) 1-hexanol; (2) n-pentadecane; (3) benzaldehyde; (4) eugenol; (5) eugenolmethylether; (6) citronellol; (7) limonene; (8) linalool; (9) β-caryophyllene; (10) trans-farnesol; (11) nerolidol.

Table 1.

Results of the ANOVA for the bioassay with ants and bumblebees

Parameter	d.f.	F	p
Insect × Substance × Concentration	4	0.16	0.96
Insect × Substance	10	20.44	<0.001
Insect × Concentration	1	1.02	0.31
Substance × Concentration	4	0.80	0.52
Insect	1	175.65	<0.001
Substance	10	18.97	<0.001
Concentration	1	0.29	0.59
Residuals	280

Significant results are highlighted in bold.

Overall, results of the bioassay presented here are consistent with those of the meta-analysis:⁴ facultative flower visitors are repelled by most floral scent compounds, while obligate flower visitors are rather attracted by the same substances, albeit not significant in the bioassay. Furthermore, within the substances that derive from the same biosynthetical pathway (benzenoids, mono- and sesquiterpenoids) those compounds possessing functional groups were stronger ant-repellent than those without functional groups. Although the meta-analysis suggests that benzenoids have no defensive properties, eugenol and eugenolmethylether had the strongest ant-repellent effect, which may be explained by the functional groups of the compounds.

Are Floral Scents Primarily Defensive Traits?

The stability of mutualisms is ensured if the costs for each of the partners are compensated by the benefits they gain from the partnership. Immobile flowers may utilise defensive secondary metabolites in combination with morphological barriers to protect themselves against an excessive exploitation by both mutualists and antagonists. Well-adapted, obligate flower visitors may have evolved the ability to cope with these defences, analogously to specialised herbivores that tolerate the specific defences of their host-plants.¹¹ In contrast to herbivory, the obligate use of floral resources is often mutually beneficial for animals and plants. This mutualism is the result of a long co-evolution, the earliest interactions between flowers and animals may have been generally detrimental for the plants' reproduction.^5,12 Thus, floral scents may have primarily served as defensive traits and were secondarily used as host-finding signals.¹² Some floral volatiles are emitted by archaic and modern as well as by insect-pollinated and non-insect pollinated angiosperms^13–15 which supports the defensive origin of floral scents. Our results suggest that attractive as well as defensive properties of floral volatiles are equally important in shaping the floral visitor spectrum of angiosperms.

Addendum to: Junker RR, Blüthgen N. Floral scents repel facultative flower visitors, but attract obligate ones. Ann Bot. 2010;105:777–782. doi: 10.1093/aob/mcq045.