Visual detection of diminutive floral guides in the bumblebee <Emphasis Type="Italic">Bombus terrestris</Emphasis> and in the honeybee <Emphasis Type="Italic">Apis mellifera</Emphasis>

Many flowers display colour patterns comprising a large peripheral colour area that serves to attract flower visitors from some distance, and a small central, contrastingly coloured area made up by stamens or floral guides. In this study, we scaled down the size of floral guides to detect the minimal size bumblebees (Bombus terrestris) and honeybees (Apis mellifera) require for guidance. We analyzed the approach and the precise contact of the antennal tips with the floral guide of artificial flowers which precedes landing and inspection. Both bumblebees and honeybees were able to make antennal contact with circular floral guides which were 2 mm in diameter; bumblebees performed better than honeybees and antennated also at floral guides smaller than 2 mm. In discrimination experiments with bumblebees, a minimum floral guide size of 2 mm was required for discrimination between artificial flowers with and without floral guides. With increasing experience bumblebees targeted close to the site of reward instead of making antennal contact with the floral guide, whereas honeybees did not alter their initial behaviour with growing experience. Bumblebees and honeybees spontaneously target diminutive floral guides to achieve physical contact with flowers by means of their antennae which helps them to inspect flowers.     

Unidirectionality of floral colour changes

Many angiosperms have arranged their flowers in inflorescences forming a distinct signalling unit to flower visitors. In some species, the flowers of inflorescences undergo a temporal colour change corresponding exactly to a change in the reward status. Based on information obtained from the spectral reflection curves of pre-change and postchage colours of flower corollas and/or floral guides, it was possible to demonstrate that the colour phase associated with reward closely corresponds to the visual stimuli which trigger behavioural responses of inexperienced flower visitors, and that the colour phase associated with less reward corresponds to visual stimuli less attractive to naïve flower visitors. Reciprocal colour changes were not observed. It is to be assumed that the unidirectionality of floral colour changes is an adaptation of angiosperms aimed at the guidance of first-time flower visitors. Signalling reward to inexperienced flower visitors is an additional function of floral colour changes. The main function of floral colour changes, however, is to provide cues with which the flower visitors can learn to associate one colour phase with reward.     

A bee’s eye view of remarkable floral colour patterns in the south-west Australian biodiversity hotspot revealed by false colour photography

Background and AimsColour pattern is a key cue of bee attraction selectively driving the appeal of pollinators. It comprises the main colour of the flower with extra fine patterns, indicating a reward focal point such as nectar, nectaries, pollen, stamens and floral guides. Such advertising of floral traits guides visitation by the insects, ensuring precision in pollen gathering and deposition. The study, focused in the Southwest Australian Floristic Region, aimed to spot bee colour patterns that are usual and unusual, missing, accomplished by mimicry of pollen and anthers, and overlapping between mimic-model species in floral mimicry cases.MethodsFloral colour patterns were examined by false colour photography in 55 flower species of multiple highly diverse natural plant communities in south-west Australia. False colour photography is a method to transform a UV photograph and a colour photograph into a false colour photograph based on the trichromatic vision of bees. This method is particularly effective for rapid screening of large numbers of flowers for the presence of fine-scale bee-sensitive structures and surface roughness that are not detectable using standard spectrophotometry.Key ResultsBee- and bird-pollinated flowers showed the expected but also some remarkable and unusual previously undetected floral colour pattern syndromes. Typical colour patterns include cases of pollen and flower mimicry and UV-absorbing targets. Among the atypical floral colour patterns are unusual white and UV-reflecting flowers of bee-pollinated plants, bicoloured floral guides, consistently occurring in Fabaceae spp., and flowers displaying a selective attractiveness to birds only. In the orchid genera (Diuris and Thelymitra) that employ floral mimicry of model species, we revealed a surprising mimicry phenomenon of anthers mimicked in turn by model species.ConclusionThe study demonstrates the applicability of ‘bee view’ colour imaging for deciphering pollinator cues in a biodiverse flora with potential to be applied to other eco regions. The technique provides an exciting opportunity for indexing floral traits on a biome scale to establish pollination drivers of ecological and evolutionary relevance.     

Floral colour change in Pedicularis monbeigiana (Orobanchaceae)

We examined the effects of the retention of colour-changed flowers on long- and short-distance attractiveness of bumblebees and the likelihood of successive flower visits by bumblebees in Pedicularis monbeigiana. The lower lip changed colour with age from white to purple. Hand geitonogamous pollination significantly reduced seed production. No pollen limitation occurred in this species. Purple-phase flowers contributed minimally to pollinator attractiveness at long distance. The combination of less reproductive flowers with a lower amount of reward and floral colour change enabled plants to direct pollinators to reproductive, highly rewarding white flowers at close range. A high percentage of purple-phase flowers in an inflorescence was associated with a marked reduction in the frequency of successive flower visits to individual plants. We suggest floral colour change in P. monbeigiana may serve as a mechanism for enhancing inter-individual pollen transfer and reducing intra-individual pollen transfer.     

A generalised mimicry system involving angiosperm flower colour, pollen and bumblebees’ innate colour preferences

Flower colour is a major advertisement signal of zoophilous plants for pollinators. Bees, the main pollinators, exhibit innate colour preferences, which have often been attributed to only one single floral colour, though most flowers display a pattern of two or several colours. The existing studies of floral colour patterns are mostly qualitative studies. Using a model of bee colour vision we quantitatively investigate two questions: whether or not component colours of floral colour patterns may mimic pollen signals, and whether or not bumblebees exhibit innate preferences for distinct parameters of naturally existing floral colour patterns. We analysed the spectral reflectances of 162 plant species with multicoloured flowers and inflorescences, distiniguishing between inner and outer colours of floral colour patterns irrespective of the particular structures so coloured.We found that:– The inner colour of radially symmetrical flowers and inflorescences and of zygomorphic flowers appears less diverse to bees than the peripheral colour.– The inner colour of most radial flowers and inflorescences as well as the inner colour of a large number of non-related zygomorphic flowers appears to bees to be very similar to that of pollen.– Bumblebees (Bombus terrestris) exhibit innate preferences for two-coloured over single-coloured dummy flowers in a spontaneous choice test.– Bumblebees exhibit innate preferences for dummy flowers with a large over those with a small centre area.– Bumblebees exhibit innate preferences for dummy flowers with a centre colour similar to that of pollen over those with another centre colour.Our findings support the hypotheses that the inner component of floral colour patterns could be interpreted as a generalised and little recognised form of mimicry of the colour of visually displayed pollen, that bumblebees exhibit innate preferences regarding colour and size parameters of floral colour patterns, and that these correspond to visually displayed pollen. These findings together suggest a prominent role of floral colour patterns in advertisement to and guidance of naive flower visitors.     

Innate Flower Recognition in Bumblebees (Bombus terrestris,B. lucorum; Apidae): Optical Signals from Stamens as Landing Reaction Releasers

Innate behavioural reactions in bumblebees were observed in flower dummy experiments. In a previous paper, it was shown that the approach of flower-naive, untrained bumblebees toward flower dummies is released by colour signals, i.e. a gradient of spectral purity. The bumblebees are directed toward that part of a flower dummy which exhibits the highest spectral purity (Lunau 1990). In the present work it is shown that the bumblebees, while still in flight, make contact with this part using their antennal tips. The antennae reaction is followed by a landing reaction which is released by olfactory and optical signals. The specific optical signals of the guide area which release the landing reaction were investigated. Since in the flower dummies which were used, the guide had the highest spectral purity, most of the antennae reactions were executed at the guide. The antennae reaction allows for a precise orientation and may involve the stimulation of a specialized area of ommatidia. Minute changes in optical guide parameters could thus be detected. The following features of stamens which most effectively released the landing reaction were: a minimum of 2 theca dummies; an oval shape of theca dummies which face each other along the longitudinal axis (both these characteristics are found in genuine stamens); 4 mm distance between theca dummies (this coincides with the distance between the antennal tips of bumblebees); two equally sized theca dummies each with an area of 7 mm² (this size of theca dummies was preferred over larger or smaller ones; the latter parameter was not changed when the corolla size was varied). Colour parameters are not involved in releasing the landing reaction. Clearly, bumblebees possess an innate releasing mechanism (IRM) that responds to optical signals of anthers which release the landing reaction. The experimental results are in accordance with conclusions drawn from the comparative studies of Osche (1979, 1983a, 1983b) that many flower guides can be interpreted as stamen mimics. Therefore stamens as well as stamen mimics represent a highly standardized signal for the insect pollinators and they are well suited for innate orientation at flowers.     

Selection of trait combinations through bee and fly visitation to flowers of Polemonium foliosissimum

Pollinators are known to exert natural selection on floral traits, but the extent to which combinations of floral traits are subject to correlational selection (nonadditive effects of two traits on fitness) is not well understood. Over two years, we used phenotypic manipulations of plant traits to test for effects of flower colour, flower shape and their interaction on rates of pollinator visitation to Polemonium foliosissimum. We also tested for correlational selection based on weighting visitation by the amount of conspecific pollen delivered per visit by each category of insect visitor. Although bumblebees were the presumed pollinators, solitary bees and flies contributed substantially (42%) to pollination. In manipulations of one trait at a time, insects visited flowers presenting the natural colour and shape over flowers manipulated to present artificial mutants with either paler colour or a more open or more tubular flower. When both colour and shape were manipulated in combination, selection on both traits arose, with bumblebees responding mainly to colour and flies responding mainly to shape. Despite selection on both floral traits, in a year with many bumblebees, we saw no evidence for correlational selection of these traits. In a year when flies predominated, fly visitation showed a pattern of correlational selection, but not favouring the natural phenotype, and correlational selection was still not detected for expected pollen receipt. These results show that flower colour and shape are subject to pollinator‐mediated selection and that correlational selection can be generated based on pollinator visitation alone, but provide no evidence for correlational selection specifically for the current phenotype.     

Colour saturation triggers innate reactions to flower signals: Flower dummy experiments with bumblebees

Summary Innate behavioural reactions, i.e. reactions of untrained, flower-naive bumblebees (Bombus terrestris L., B. lucorum L.; Apidae) were observed in flower dummy experiments. It was proven that an innate releasing mechanism responds to optical flower signals: the spectral purity of corolla colour was found to be crucial for far attraction toward flower dummies. During the subsequent near orientation, that is when a bumblebee finally reaches a flower dummy, the bumblebee's antennae contact the part of highest spectral purity while the bee is still in flight. Guides such as stamen patches present in the center of flower dummies are used only for near orientation. Flower dummies receiving the greatest number of antennae reactions at the guide were always those with low spectral purity in the surrounding background colour, high spectral purity at the corolla colour and highest spectral purity at the guide colour. In contrast, dominant wavelength and intensity of flower dummy colours had no detectable influence on innate behavioural reactions, while colour contrast had some. These results are interpreted as follows: orientation toward guides is based upon a gradient of centripetally increasing, bee-subjective colour saturation which directs the bumblebee's approach toward the center of the flower dummy where additional factors may contribute to stimulating the landing reaction.     

Bumblebees (Bombus terrestris) and honeybees (Apis mellifera) prefer similar colours of higher spectral purity over trained colours

Differences in the concentration of pigments as well as their composition and spatial arrangement cause intraspecific variation in the spectral signature of flowers. Known colour preferences and requirements for flower-constant foraging bees predict different responses to colour variability. In experimental settings, we simulated small variations of unicoloured petals and variations in the spatial arrangement of colours within tricoloured petals using artificial flowers and studied their impact on the colour choices of bumblebees and honeybees. Workers were trained to artificial flowers of a given colour and then given the simultaneous choice between three test colours: either the training colour, one colour of lower and one of higher spectral purity, or the training colour, one colour of lower and one of higher dominant wavelength; in all cases the perceptual contrast between the training colour and the additional test colours was similarly small. Bees preferred artificial test flowers which resembled the training colour with the exception that they preferred test colours with higher spectral purity over trained colours. Testing the behaviour of bees at artificial flowers displaying a centripetal or centrifugal arrangement of three equally sized colours with small differences in spectral purity, bees did not prefer any type of artificial flowers, but preferentially choose the most spectrally pure area for the first antenna contact at both types of artificial flowers. Our results indicate that innate preferences for flower colours of high spectral purity in pollinators might exert selective pressure on the evolution of flower colours.     

Frequency-dependent selection by pollinators: mechanisms and consequences with regard to behaviour of bumblebees Bombus terrestris (L.) (Hymenoptera: Apidae)

Bombus terrestris , a typical pollinating insect species, was offered artificial flowers of two different corolla colours with the same sucrose solution reward in an array. Common colours were significantly preferred, and the strength of the frequency-dependent response increased as a result of learning. There were also frequency-independent biases towards blue flowers, probably because blue flowers appeared more conspicuous to bumblebees than yellow flowers, and the degree of preference for blue was greater when flowers had low nectar rewards. Flower-to-flower movements by individual bumblebees between flowers were non-random, were biased to movements within the same flower colour, and were also dependent on morph frequency. The mechanisms governing flower selection in bumblebees are discussed. Pollinators foraging similarly in a natural situation would induce positive frequency-dependent selection, assortative mating, and directional selection on different corolla colour morphs of the plant population being visited, resulting in stabilizing selection for a single flower colour.     

Density-dependent and frequency-dependent selection by bumblebees Bombus terrestris (L.) (Hymenoptera: Apidae)

The behaviour of bumblebee workers foraging on arrays of artificial flowers of two colour morphs was observed. Experiments were conducted on arrays of varying morph frequencies and at three different total flower densities. Bumblebees consistently showed a preference for the commonest colour morph, and this behaviour was not significantly affected by changing density. In contrast, frequency-independent preferences changed significantly with density. At low densities, there was a strong bias towards the more conspicuous colour, whereas at higher densities there was no overall colour bias. Flight distances between flowers decreased significantly at high density. Bumblebees also visited flowers of similar colours sequentially, but this behaviour was not density-dependent. It is suggested that as densities increase, there is an increased probability that bumblebees detect yellow flowers, which were probably less conspicuous compared with blue flowers, and that this might be caused by changes in flight speed with flight distance. Where there is a positive relationship between pollinator visitation and the relative fitness of a floral morph, the observed behaviour would induce positive frequency-dependent selection on a plant population with two corolla colour morphs on which the bumblebees were foraging, which would result in stabilizing selection for a single corolla colour, irrespective of density. There was no indication that rare colour morphs would be preferred at high density. The probability of different corolla colour morphs going to fixation would, however, be affected by density.     

A Matter of Contrast: Yellow Flower Colour Constrains Style Length in Crocus species

Most flowers display distinct colour patterns comprising two different areas. The peripheral large-area component of floral colour patterns attracts flower visitors from some distance and the central small-area component guides flower visitors towards landing sites. Whereas the peripheral colour is largely variable among species, the central colour, produced mostly by anthers and pollen or pollen mimicking floral guides, is predominantly yellow and UV-absorbing. This holds also for yellow flowers that regularly display a UV bull’s eye pattern. Here we show that yellow-flowering Crocus species are a noticeable exception, since yellow-flowering Crocus species–being entirely UV-absorbing–exhibit low colour contrast between yellow reproductive organs and yellow tepals. The elongated yellow or orange-yellow style of Crocus flowers is a stamen-mimicking structure promoting cross-pollination by facilitating flower visitors’ contact with the apical stigma before the flower visitors are touching the anthers. Since Crocus species possess either yellow, violet or white tepals, the colour contrast between the stamen-mimicking style and the tepals varies among species. In this study comprising 106 Crocus species, it was tested whether the style length of Crocus flowers is dependent on the corolla colour. The results show that members of the genus Crocus with yellow tepals have evolved independently up to twelve times in the genus Crocus and that yellow-flowering Crocus species possess shorter styles as compared to violet- and white-flowering ones. The manipulation of flower visitors by anther-mimicking elongated styles in Crocus flowers is discussed.     

Flower colour adaptation in a mimetic orchid

Although the tremendous variability in floral colour among angiosperms is often attributed to divergent selection by pollinators, it is usually difficult to preclude the possibility that floral colour shifts were driven by non-pollinator processes. Here, we examine the adaptive significance of flower colour in Disa ferruginea, a non-rewarding orchid that is thought to attract its butterfly pollinator by mimicking the flowers of sympatric nectar-producing species. Disa ferruginea has red flowers in the western part of its range and orange flowers in the eastern part--a colour shift that we hypothesized to be the outcome of selection for resemblance to different local nectar-producing plants. Using reciprocal translocations of red and orange phenotypes as well as arrays of artificial flowers, we found that the butterfly Aeropetes tulbaghia, the only pollinator of the orchid, preferred both the red phenotype and red artificial flowers in the west where its main nectar plant also has red flowers, and both the orange phenotype and orange artificial flowers in the east, where its main nectar plant has orange flowers. This phenotype by environment interaction demonstrates that the flower colour shift in D. ferruginea is adaptive and driven by local colour preference in its pollinator.     

Colour choices of naive bumble bees and their implications for colour perception

The innate preferences of inexperienced bumble bees, Bombus terrestris, for floral colour stimuli were studied using artificial flowers. The artificial flowers provided a colour pattern and consisted of a star-shaped corolla and of central colour patches similar to the nectar guide of natural flowers. The innate choice behaviour was assessed in terms of the number of approach flights from some distance towards the artificial flowers and the percentage of approach flights terminating in antennal contact with the floral guide. The colours of the floral guide, the corolla and the background were varied. It was shown that the innate flower colour preference in bumble bees has two components. 1. The frequency of approaches from a distance is correlated with the colour difference between the corolla and the background against which it is presented. If the corolla colour was constant but its background colour varied, the relative attractiveness of the corolla increased with its colour difference to the background. The colour difference assessment underlying this behaviour on a perceptual basis can be attained by means of colour opponent coding, a system well-established in Hymenoptera. 2. The frequency of antennal contacts with the floral guides relative to that of approach flights cannot be accounted for by colour opponent coding alone. Whether the approach flights are interrupted, or whether they end in an antennal contact with the nectar guide is strongly dependent on the direction (sign) of the colour difference, not only its magnitude. The choice behaviour requires a unique perceptual dimension, possibly that of colour saturation or that of hue perception comparable to components of colour perception in humans.     

The movement patterns of carpenter beesXylocopa micans and bumblebeesBombus pennsylvanicus onPontederia cordata inflorescences

The movement patterns of carpenter bees (Xylocopa micans) and bumblebees (Bombus pennsylvanicus) foraging for nectar on vertical inflorescences ofPontederia cordata were studied near Miami, Florida. The floral biology ofP. cordata is unique in several ways: (a) many short-lived flowers per inflorescence, (b) constant nectar production throughout the life span of each flower, and (c) abscence of vertical patterning of nectar and age of flowers. Inflorescences ranged between 3.5 and 15.8 cm long and had between 9 and 55 open flowers. Both carpenter bees and bumblebees arrived mostly on the bottom third of the inflorescence and left after visiting flowers on the top third of the inflorescence. The departure position from the inflorescence was higher up than observed in studies of other insect pollinators foraging on other speces of plants. This pattern of departure probably occurs in the absence of a vertical gradient of nectar or floral morphology.     

Flower colours along an alpine altitude gradient, seen through the eyes of fly and bee pollinators

Alpine flowers face multiple challenges in terms of abiotic and biotic factors, some of which may result in selection for certain colours at increasing altitude, in particular the changing pollinator species composition, which tends to move from bee-dominated at lower elevations to fly-dominated in high-alpine regions. To evaluate whether growing at altitude—and the associated change in the dominant pollinator groups present—has an effect on the colour of flowers, we analysed data collected from the Dovrefjell National Park in Norway. Unlike previous studies, however, we considered the flower colours according to ecologically relevant models of bee and fly colour vision and also their physical spectral properties independently of any colour vision system, rather than merely looking at human colour categories. The shift from bee to fly pollination with elevation might, according to the pollination syndrome hypothesis, lead to the prediction that flower colours should shift from more bee-blue and UV-blue flowers (blue/violet to humans, i.e. colours traditionally associated with large bee pollinators) at low elevations to more bee-blue-green and green (yellow and white to humans—colours often linked to fly pollination) flowers at higher altitude. However, although there was a slight increase in bee-blue-green flowers and a decrease in bee-blue flowers with increasing elevation, there were no statistically significant effects of altitude on flower colour as seen either by bees or by flies. Although flower colour is known to be constrained by evolutionary history, in this sample we also did not find evidence that phylogeny and elevation interact to determine flower colours in alpine areas. Handling editor: Neal Williams     

Pollen and stamen mimicry: the alpine flora as a case study

Many melittophilous flowers display yellow and UV-absorbing floral guides that resemble the most common colour of pollen and anthers. The yellow coloured anthers and pollen and the similarly coloured flower guides are described as key features of a pollen and stamen mimicry system. In this study, we investigated the entire angiosperm flora of the Alps with regard to visually displayed pollen and floral guides. All species were checked for the presence of pollen- and stamen-imitating structures using colour photographs. Most flowering plants of the Alps display yellow pollen and at least 28% of the species display pollen- or stamen-imitating structures. The most frequent types of pollen and stamen imitations were (mostly yellow and UV-absorbing) colour patches on petals (65% of species displaying imitations), patterns of inflorescences (18%), stamen-like pistils (10%), and staminodes (6%), as well as three-dimensional structures such as convex lower lips and filamental hairs (<5%). Dichogamous and diclinous species display pollen- and stamen-imitating structures more often than non-dichogamous and non-diclinous species, respectively. The visual similarity between the androecium and other floral organs is attributed to mimicry, i.e. deception caused by the flower visitor’s inability to discriminate between model and mimic, sensory exploitation, and signal standardisation among floral morphs, flowering phases, and co-flowering species. We critically discuss deviant pollen and stamen mimicry concepts and evaluate the frequent evolution of pollen-imitating structures in view of the conflicting use of pollen for pollination in flowering plants and provision of pollen for offspring in bees.     

Colour evolution within orchids depends on whether the pollinator is a bee or a fly

• Orchids are a classic angiosperm model for understanding biotic pollination. We studied orchid species within two species‐rich herbaceous communities that are known to have either hymenopteran or dipteran insects as the dominant pollinators, in order to understand how flower colour relates to pollinator visual systems.
• We analysed features of the floral reflectance spectra that are significant to pollinator visual systems and used models of dipteran and hymenopteran colour vision to characterise the chromatic signals used by fly‐pollinated and bee‐pollinated orchid species.
• In contrast to bee‐pollinated flowers, fly‐pollinated flowers had distinctive points of rapid reflectance change at long wavelengths and a complete absence of such spectral features at short wavelengths. Fly‐pollinated flowers also had significantly more restricted loci than bee‐pollinated flowers in colour space models of fly and bee vision alike.
• Globally, bee‐pollinated flowers are known to have distinctive, consistent colour signals. Our findings of different signals for fly pollination is consistent with pollinator‐mediated selection on orchid species that results from the distinctive features of fly visual systems.

     

Consequences of floral complexity for bumblebee-mediated geitonogamous self-pollination in Salvia nipponica Miq. (Labiatae)

Abstract.— I address how floral complexity influences geitonogamous self-pollination through manipulation of pollinator behavior in Salvia nipponica . The pivoting stamens of S. nipponica hinder nectar-collecting bumblebees from crawling into flowers, increasing the probing time per flower. I predicted that longer probing times would reduce the relative cost of moving between plants, causing bees to leave plants earlier. To test this prediction, I simplified S. nipponica flowers by removing the stamens from all open flowers within a 75-m² quadrat. Bumblebees probed these flowers more quickly than intact flowers, but the stamen removal affected neither the frequency of flower revisitation nor the flight distance between plants. In response to the decrease in the probing time per flower, bees probed more flowers on these plants. Therefore, in S. nipponica , floral complexity reduces the opportunity for geitonogamous self-pollination. Stamen removal also increased bee visitation per flower, suggesting that this sort of complexity deters visitation. To keep complex flowers attractive, therefore, selection might increase floral rewards or longevity. Floral complexity might evolve in an integrative manner with the rest of the floral phenotype.     

Gloss,Colour and Grip: Multifunctional Epidermal Cell Shapes in Bee- and Bird-Pollinated Flowers

Flowers bear the function of filters supporting the attraction of pollinators as well as the deterrence of floral antagonists. The effect of epidermal cell shape on the visual display and tactile properties of flowers has been evaluated only recently. In this study we quantitatively measured epidermal cell shape, gloss and spectral reflectance of flowers pollinated by either bees or birds testing three hypotheses: The first two hypotheses imply that bee-pollinated flowers might benefit from rough surfaces on visually-active parts produced by conical epidermal cells, as they may enhance the colour signal of flowers as well as the grip on flowers for bees. In contrast, bird-pollinated flowers might benefit from flat surfaces produced by flat epidermal cells, by avoiding frequent visitation from non-pollinating bees due to a reduced colour signal, as birds do not rely on specific colour parameters while foraging. Moreover, flat petal surfaces in bird-pollinated flowers may hamper grip for bees that do not touch anthers and stigmas while consuming nectar and thus, are considered as nectar thieves. Beside this, the third hypothesis implies that those flower parts which are vulnerable to nectar robbing of bee- as well as bird-pollinated flowers benefit from flat epidermal cells, hampering grip for nectar robbing bees. Our comparative data show in fact that conical epidermal cells are restricted to visually-active parts of bee-pollinated flowers, whereas robbing-sensitive parts of bee-pollinated as well as the entire floral surface of bird-pollinated flowers possess on average flat epidermal cells. However, direct correlations between epidermal cell shape and colour parameters have not been found. Our results together with published experimental studies show that epidermal cell shape as a largely neglected flower trait might act as an important feature in pollinator attraction and avoidance of antagonists, and thus may contribute to the partitioning of flower-visitors.