Multimodal cues provide redundant information for bumblebees when the stimulus is visually salient, but facilitate red target detection in a naturalistic background

Floral volatiles evoke partially similar responses in both florivores and pollinators and are correlated with non-volatile reward chemicals

BACKGROUND

Plants often use floral displays to attract mutualists and prevent antagonist attacks. Chemical displays detectable from a distance include attractive or repellent floral volatile organic compounds (FVOCs). Locally, visitors perceive contact chemicals including nutrients but also deterrent or toxic constituents of pollen and nectar. The FVOC and pollen chemical composition can vary intra- and interspecifically. For certain pollinator and florivore species, responses to these compounds are studied in specific plant systems, yet we lack a synthesis of general patterns comparing these two groups and insights into potential correlations between FVOC and pollen chemodiversity.

SCOPE

We reviewed how FVOCs and non-volatile floral chemical displays, i.e. pollen nutrients and toxins, vary in composition and affect the detection by and behaviour of insect visitors. Moreover, we used meta-analyses to evaluate the detection of and responses to FVOCs by pollinators vs. florivores within the same plant genera. We also tested whether the chemodiversity of FVOCs, pollen nutrients and toxins is correlated, hence mutually informative.

KEY RESULTS

According to available data, florivores could detect more FVOCs than pollinators. Frequently tested FVOCs were often reported as pollinator-attractive and florivore-repellent. Among FVOCs tested on both visitor groups, there was a higher number of attractive than repellent compounds. FVOC and pollen toxin richness were negatively correlated, indicating trade-offs, whereas a marginal positive correlation between the amount of pollen protein and toxin richness was observed.

CONCLUSIONS

Plants face critical trade-offs, because floral chemicals mediate similar information to both mutualists and antagonists, particularly through attractive FVOCs, with fewer repellent FVOCs. Furthermore, florivores might detect more FVOCs, whose richness is correlated with the chemical richness of rewards. Chemodiversity of FVOCs is potentially informative of reward traits. To gain a better understanding of the ecological processes shaping floral chemical displays, more research is needed on floral antagonists of diverse plant species and on the role of floral chemodiversity in visitor responses.

Multimodal floral recognition by bumblebees

Flowers present information to their insect visitors in multiple simultaneous sensory modalities. Research has commonly focussed on information presented in visual and olfactory modalities. Recently, focus has shifted towards additional 'invisible' information, and whether information presented in multiple modalities enhances the interaction between flowers and their visitors. In this review, we highlight work that addresses how multimodality influences behaviour, focussing on work conducted on bumblebees (Bombus spp.), which are often used due to both their learning abilities and their ability to use multiple sensory modes to identify and differentiate between flowers. We review the evidence for bumblebees being able to use humidity, electrical potential, surface texture and temperature as additional modalities, and consider how multimodality enhances their performance. We consider mechanisms, including the cross-modal transfer of learning that occurs when bees are able to transfer patterns learnt in one modality to an additional modality without additional learning.

Contrasting effects of fungicide and herbicide active ingredients and their formulations on bumblebee learning and behaviour

Fungicides and herbicides are two of the most heavily applied pesticide classes in the world, but receive little research attention with regards to their potential impacts on bees. As they are not designed to target insects, the mechanisms behind potential impacts of these pesticides are unclear. It is therefore important to understand their influence at a range of levels, including sublethal impacts on behaviours such as learning. We used the proboscis extension reflex (PER) paradigm to assess how the herbicide glyphosate and the fungicide prothioconazole affect bumblebee olfactory learning. We also assessed responsiveness, and compared the impacts of these active ingredients and their respective commercial formulations (Roundup Biactive and Proline). We found that learning was not impaired by either formulation but, of the bees that displayed evidence of learning, exposure to prothioconazole active ingredient increased learning level in some situations, while exposure to glyphosate active ingredient resulted in bumblebees being less likely to respond to antennal stimulation with sucrose. Our data suggest that fungicides and herbicides may not negatively impact olfactory learning ability when bumblebees are exposed orally to field-realistic doses in a lab setting, but that glyphosate has the potential to cause changes in responsiveness in bees. As we found impacts of active ingredients and not commercial formulations, this suggests that co-formulants may modify impacts of active ingredients in the products tested on olfactory learning without being toxic themselves. More research is needed to understand the mechanisms behind potential impacts of fungicides and herbicides on bees, and to evaluate the implications of behavioural changes caused by glyphosate and prothioconazole for bumblebee fitness.

A sophisticated case of division of labour in the trimorphic stamens of the Cassia fistula (Leguminosae) flower

Buzz-pollinated pollen flowers have pollen as the primary resource for pollinators and must deal with a conflict between the exploitation of pollen grains by bees and pollination success. It has been hypothesized that heterostemony allows division of labour between stamens as a solution to the pollen dilemma. To test the division of labour hypothesis, we chose Cassia fistula, which has a trimorphic androecium and analysed androecium development, pollen grain release mechanisms and visitor behaviour. We explored the reflectance of floral organs and carried out an exclusion experiment to test the attractiveness of each stamen morph to the bee species. Finally, we explored the structural, ultrastructural and functional variation between the pollen grains, including pollen viability across stamen morphs. The differences among the three stamen morphs, which is developed from two whorls of the stamen, are the first evidence of the division of labour in our study system. Large Bombus and Xylocopa bees actively and exclusively exploited the pollen grains from the central poricidal anthers generating pollen deposition on their bodies. The reflectance pattern of floral organs indicated a targeting of these large bees to the central anthers, corroborated by the anther manipulative experiment where only the exclusion of the anthers positioned in the flower centre, especially the intermediate stamens, reduced bee visits. Both results revealed a division of labour, in which the intermediate stamen morph was responsible for both floral attractiveness and pollen resources. Only the largest stamen morph produced germinable pollen grains, highlighting their role as pollinating stamens. The smallest stamen morph has a less clear function, likely representing an economy in pollen production for feeding function. Our findings suggest that the evolution of the trimorphic androecium is associated with division of labour in large pollen flowers and can represent a strong strategy for circumventing the pollen dilemma, optimizing the feeding function by reducing pollen grain investment from central anthers.

Bumble bees show an induced preference for flowers when primed with caffeinated nectar and a target floral odor

Caffeine is a widely occurring plant defense chemical^1,2 that occurs in the nectar of some plants, e.g., Coffea or Citrus spp., where it may influence pollinator behavior to enhance pollination.^3,4 Honey bees fed caffeine form longer lasting olfactory memory associations,⁵ which could give plants with caffeinated nectar an adaptive advantage by inducing more visits to flowers. Caffeinated free-flying bees show enhanced learning performance⁶ and are more likely to revisit a caffeinated target feeder or artificial flower,^7-9 although it is not clear whether improved memory of the target cues or the perception of caffeine as a reward is the cause. Here, we show that inexperienced bumble bees (Bombus terrestris) locate new food sources emitting a learned floral odor more consistently if they have been fed caffeine. In laboratory arena tests, we fed bees a caffeinated food alongside a floral odor blend (priming) and then used robotic experimental flowers¹⁰ to disentangle the effects of caffeine improving memory for learned food-associated cues versus caffeine as a reward. Inexperienced bees primed with caffeine made more initial visits to target robotic flowers emitting the target odor compared to control bees or those primed with odor alone. Caffeine-primed bees tended to improve their floral handling time faster. Although the effects of caffeine were short lived, we show that food-locating behaviors in free-flying bumble bees can be enhanced by caffeine provided in the nest. Consequently, there is potential to redesign commercial colonies to enhance bees' forage focus or even bias bees to forage on a specific crop.

An Innate Preference of Bumblebees for Volatile Organic Compounds Emitted by Phaseolus vulgaris Plants Infected With Three Different Viruses

Cucumber mosaic virus (CMV)-infected tomato (Solanum lycopersicum L.) plants emit volatile organic compounds (VOCs) attractive to bumblebees (Bombus terrestris L.), which are important tomato pollinators, but which do not transmit CMV. We investigated if this effect was unique to the tomato-CMV pathosystem. In two bean (Phaseolus vulgaris L.) cultivars, infection with the potyviruses bean common mosaic virus (BCMV) or bean common mosaic necrosis virus (BCMNV), or with the cucumovirus CMV induced quantitative changes in VOC emission detectable by coupled gas chromatography–mass spectrometry. In free-choice olfactometry assays bumblebees showed an innate preference for VOC blends emitted by virus-infected non-flowering bean plants and flowering CMV-infected bean plants, over VOCs emitted by non-infected plants. Bumblebees also preferred VOCs of flowering BCMV-infected plants of the Wairimu cultivar over non-infected plants, but the preference was not significant for BCMV-infected plants of the Dubbele witte cultivar. Bumblebees did not show a significant preference for VOCs from BCMNV-infected flowering bean plants but differential conditioning olfactometric assays showed that bumblebees do perceive differences between VOC blends emitted by flowering BCMNV-infected plants over non-infected plants. These results are consistent with the concept that increased pollinator attraction may be a virus-to-host payback, and show that virus-induced changes in bee-attracting VOC emission is not unique to one virus-host combination.

Humans Share More Preferences for Floral Phenotypes With Pollinators Than With Pests

Studies on the selection of floral traits usually consider pollinators and sometimes herbivores. However, humans also exert selection on floral traits of ornamental plants. We compared the preferences of bumblebees (Bombus terrestris), thrips (Frankliniella occidentalis), and humans for flowers of snapdragon. From a cross of two species, Antirrhinum majus and Antirrhinum linkianum, we selected four Recombinant Inbred Lines (RILs). We characterised scent emission from whole flowers and stamens, pollen content and viability, trichome density, floral shape, size and colour of floral parts. We tested the preferences of bumblebees, thrips, and humans for whole flowers, floral scent bouquets, stamen scent, and individual scent compounds. Humans and bumblebees showed preferences for parental species, whereas thrips preferred RILs. Colour and floral scent, in combination with other floral traits, seem relevant phenotypes for all organisms. Remarkably, visual traits override scent cues for bumblebees, although, scent is an important trait when bumblebees cannot see the flowers, and methyl benzoate was identified as a key attractant for them. The evolutionary trajectory of flowers is the result of multiple floral traits interacting with different organisms with different habits and modes of interaction.

Flowering Phenology and the Influence of Seasonality in Flower Conspicuousness for Bees

Flowering patterns are crucial to understand the dynamics of plant reproduction and resource availability for pollinators. Seasonal climate constrains flower and leaf phenology, where leaf and flower colors likely differ between seasons. Color is the main floral trait attracting pollinators; however, seasonal changes in the leaf-background coloration affect the perception of flower color contrasts by pollinators. For a seasonally dry woody cerrado community (Brazilian savanna) mainly pollinated by bees, we verified whether seasonality affects flower color diversity over time and if flower color contrasts of bee-pollinated species differ between seasons due to changes in the leaf-background coloration. For 140 species, we classified flower colors based on human-color vision, and for 99 species, we classified flower colors based on bee-color vision (spectral measurements). We described the community's flowering pattern according to the flower colors using a unique 11 years phenological database. For the 43 bee-pollinated species in which reflectance data were also available, we compared flower color diversity and contrasts against the background between seasons, considering the background coloration of each season. Flowering was markedly seasonal, peaking at the end of the dry season (September), when the highest diversity of flower colors was observed. Yellow flowers were observed all year round, whereas white flowers were seasonal, peaking during the dry season, and pink flowers predominated in the wet season, peaking in March. Bee-bluegreen flowers peaked between September and October. Flowers from the wet and dry seasons were similarly conspicuous against their corresponding background. Regardless of flowering season, the yellowish background of the dry season promoted higher flower color contrast for all flower species, whereas the greener background of the wet season promoted a higher green contrast. Temporal patterns of flower colors and color contrasts were related to the cerrado seasonality, but also to bee's activity, visual system, and behavior. Background coloration affected flower contrasts, favoring flower conspicuousness to bees according to the season. Thus, our results provide new insights regarding the temporal patterns of plant-pollinator interactions.

What pollinators see does not match what they smell: Absence of color-fragrance association in the deceptive orchid Ionopsis utricularioides

Many deceptive orchids present variation in floral color and fragrance. This might be advantageous for the plant, as it can disturb the associative avoidance learning of pollinators, promoting more visits to the flowers. Some studies have shown that color and fragrance can be correlated in polymorphic deceptive orchids, but these studies employed color traits based on the human visual system and not the visual perception of pollinators. Thus, we investigated the composition of the floral fragrance of Ionopsis utricularioides (Sw.) Lindl., a polymorphic deceptive orchid, and analyzed possible correlations with the floral color as seen by bees, Apis mellifera L. and Melipona quadrifasciata Lepeletier, using the color hexagon model. We found high color and fragrance intraspecific variation, as expected for deceptive species. However, we found no color-fragrance association in individuals, either by comparing fragrance profiles with the color variable saturation or by comparing them with the placement of individuals in the color hexagon for both bee species. This lack of correlation contradicts the biochemical pathway hypothesis, which proposes that associations between floral color and scent in polymorphic flowers arise from shared biochemical pathways. However, a complete absence of correlation between floral signals is consistent with selection arising through pollinator cognitive ecology. Lack of correlation would increase the floral variability perceived by bees, given their multimodal learning, and this variability could disrupt avoidance learning of deceptive flowers, thus enhancing the efficacy of the plant's deceptive pollination mechanism.

What gaze direction can tell us about cognitive processes in invertebrates

Most visually guided animals shift their gaze using body movements, eye movements, or both to gather information selectively from their environments. Psychological studies of eye movements have advanced our understanding of perceptual and cognitive processes that mediate visual attention in humans and other vertebrates. However, much less is known about how these processes operate in other organisms, particularly invertebrates. We here make the case that studies of invertebrate cognition can benefit by adding precise measures of gaze direction. To accomplish this, we briefly review the human visual attention literature and outline four research themes and several experimental paradigms that could be extended to invertebrates. We briefly review selected studies where the measurement of gaze direction in invertebrates has provided new insights, and we suggest future areas of exploration.

Unlimited Associative Learning and the origins of consciousness: a primer and some predictions

Over the past two decades, Ginsburg and Jablonka have developed a novel approach to studying the evolutionary origins of consciousness: the Unlimited Associative Learning (UAL) framework. The central idea is that there is a distinctive type of learning that can serve as a transition marker for the evolutionary transition from non-conscious to conscious life. The goal of this paper is to stimulate discussion of the framework by providing a primer on its key claims (Part I) and a clear statement of its main empirical predictions (Part II).

Red flowers differ in shades between pollination systems and across continents

BACKGROUND AND AIMS

Floral colour is a primary signal in plant-pollinator interactions. The association between red flowers and bird pollination is well known, explained by the 'bee avoidance' and 'bird attraction' hypotheses. Nevertheless, the relative importance of these two hypotheses has rarely been investigated on a large scale, even in terms of colour perception per se.

METHODS

We collected reflectance spectra for 130 red flower species from different continents and ascertained their pollination systems. The spectra were analysed using colour vision models for bees and (three types of) birds, to estimate colour perception by these pollinators. The differences in colour conspicuousness (chromatic and achromatic contrast, purity) and in spectral properties between pollination systems and across continents were analysed.

KEY RESULTS

Compared with other floral colours, red flowers are very conspicuous to birds and much less conspicuous to bees. The red flowers pollinated by bees and by birds are more conspicuous to their respective pollinators. Compared with the bird flowers in the Old World, the New World ones are less conspicuous to bees and may be more conspicuous not only to violet-sensitive but also to ultraviolet-sensitive birds. These differences can be explained by the different properties of the secondary reflectance peak (SP). SP intensity is higher in red flowers pollinated by bees than those pollinated by birds (especially New World bird flowers). A transition from high SP to low SP in red flowers can induce chromatic contrast changes, with a greater effect on reducing attraction to bees than enhancing attraction to birds.

CONCLUSIONS

Shades of red flowers differ between pollination systems. Moreover, red bird flowers are more specialized in the New World than in the Old World. The evolution towards colour specialization is more likely to result in higher efficiency of bee avoidance than bird attraction.

Flower Conspicuousness to Bees Across Pollination Systems: A Generalized Test of the Bee-Avoidance Hypothesis

Flower signals of bee- and bird-pollinated plants have converged via pollinator-mediated evolution, driven by the visual system of their respective pollinators. For bird flowers, sensory exclusion of less effective bees is also important and such exclusion is also mediated by floral morphological filters. Likewise, other systems based on pollination by red-sensitive insects are also associated with red flowers displaying lower short-wavelength secondary peaks of reflectance, which decreases detectability to animals that are less sensitive to red, such as bees. These flowers often also present long tubes. Here, we tested a generalization of the bee-avoidance hypothesis in order to assess if it holds only for bird flowers or for other non-bee pollination systems as well. For this, we compared flower contrasts and spectral purity in bee visual systems as proxies for conspicuousness among four kinds of pollination systems: bee-visited flowers, insect-visited flowers (including bees and other insects), non-bee insect flowers (flowers visited by red-sensitive insects such as flies, butterflies and beetles, but not bees), and bird-visited flowers. We also assessed the association between conspicuousness to bees and flower depth, used as a proxy for morphological exclusion of bees. Overall, flower conspicuousness to bees differed only between insect (all three groups) and bird flowers, due to lower visual signals for the latter. This suggests that bee sensory exclusion via color signals is exclusive to bird flowers, while non-bee insect flowers might use other sensory channels to exclude bees, such as olfactory signals. Visual bee avoidance might be a mechanism exclusive to plants pollinated by specific guilds of red-sensitive insects not well represented in our sample. We also found a negative association between flower conspicuousness to bees and flower depth, suggesting an interplay of morphological and spectral traits in discouraging bee visits. Our results support the bee-avoidance hypothesis exclusively for bird flowers and an overall association between lower visual signals to bees and long tubes.

Encounters with predators fail to trigger predator avoidance in bumblebees, Bombus terrestris (Hymenoptera: Apidae)

Many species must learn to identify their predators, but little is known about the effect of direct encounters on the development of predator avoidance. We asked whether bumblebees, Bombus terrestris, learn to avoid predators, whether learning depends on the conspicuousness of predators and whether bumblebees learn to identify predators or simply to avoid dangerous patches. To answer these questions, we allowed bumblebees to forage in an enclosed meadow of 15 artificial flowers containing a yellow female crab spider, Thomisus onustus. Flowers were yellow in half of the trials and white in the other half. Spiders could remain at the same flower throughout the experiment or swap flowers between bee foraging bouts. Of the 60 bees used in the experiment, eight were killed by the spiders and nine stopped foraging without finishing the trial. Death or refusal to forage typically occurred early in the trial. Regardless of the treatment, the probability of landing at the spider-harbouring flower increased with time. Previous encounters with heterospecific individuals can therefore be a poor source of information about their predatory nature.