I think I remenber they qre mostly attracted by UV coloration of flora but I guess the volatiles are also important.

Generally spoken, honeybees have innate preferences for distinct chemical cues and colour parameters. Honeybees also can learn chemical signals and colour signals of flowers. The detectability of these signals is impartant, e.g. contrast against background. Honeybees also can lean some signals better than others. Floral signals as for example floral guide can redice the handling time of flowers and thus are attractive in terms of foraging efficiency. Limits are set by the sensorial capabilities, e.g. low sensitivity for red light. There is no evidence for a special role for ultraviolet colours; attractive are distinct colour hues, the saturation of colours and gree contrast as well as colour contrast. Flower colous as well as scents have a double role which is attction of pollinations as well as discouraging of nectar robber and pollen thieves; for example in bird-pollinated flowers the exclusion of honeybees play an important role.

Yes, they do... But you should actually start by reading some of the publications by researcher Lars Chittka, check his webpage as he has been uploading all his publications. Everything you want to know about what is attractive for bees is there.

Thank you 

Kessler et al. (2015) showed that bees (honeybees and bumblebees) cannot perceive neonics as a taste. Interestingly, they seem even to prefer nectar/sugar solution with low concentrations of neonics. The authors explain that these substances have a pharmacological action on certain receptors in the bee brain.

Kessler S.C., Tiedeken E.J., Simcock K.L., Derveau S., Mitchell J., Softley S., Stout J.C., Wright G.A. (2015) Bees prefer foods containing neonicotinoid pesticides. Nature doi: 10.1038/14414

To be sure about the meaning of attraction. Floral rewards like sugars in nectar or proteins in pollen are the ultimate causes of the attractiveness of flowers, but the the proximate causes of attraction are floral cues that flower visitors can sense. Of course sweetness of nectar correlates with the amount of sugar reward, but can only be sensed upon contact with taste receptors. This might be different with coloured visibly offered nectar indicating at leastthe amont of nectar or pollen displayed in open pollen sacs indicating honestly the amount of pollen.

Generally, it seems resonable that the preferences of floral insect visitors are multifactorial and refer to both quality and quantity of floral rewards.  For example, during my obeservations of insect visitors preferences towards plant species in ruderal phytocoenoses I found that the protein content in pollen did not affect foraging preferences. Interestingly, the presence of starch in pollen grains determines the presence of pollinators. Solitary bees, Diptera and Syrphidae eagerly visit plant species that contain starch in pollen grains, this source of energy is insignificant for Apis mellifera, whereas species of the genus Bombus avoid starchy pollen.

Apis mellifera prefers to visit plant species with an even small amount of pollen produced in flowers, but with an abundant flowering resulting in a high pollen yield per unit area over a short period of time. Other Apoidea eagerly collect pollen when the species bloom less abundantly and their pollen yield is low.

Bozenas answer raised intersting question: Are bees and hoerflies capable to sense the protein content and starch content before eating pollen and when eating pollen or when digesting pollen? Does the protein and starch content of pollen correlate with some other parameters like length of the pistil that the pollen tube has to grow through? Does the pollenkitt possess any substance indicating pollen quality; as an example see our attached ideas and findings about proline in pollenkit and hoverflies high sensitivity for this amini acid.

Dear Profesor Lunau,

Thank you very much for multiple questions you put to the discussion. It is accepted that nectar and pollen traits are plant derivative and are selected for plant reproductive success. The correlation between protein content and starch content and pollen tubes lenght and therefore the pollen potential for fertilization has been proven (e.g. Pacini, Nepi).  On the other hand, the scent released from the surface of pollen grains (from pollenkit or egzyna proteins) as suggested by Lunau [2000], Castellanos et al. [2006] and Hanley et al. [2008] . These impact on floral visitor sensors and may impact insect behavoiur. My correlations and conclusions were made based on pollinator groups frequency to different plants in ruderal biocenoses. Various plant traits (including pollen quality and quantity ) were used in multi-factorial statistical analysis. Does statistic say a truth ? Such correlations usualy do not explain all possible factors.  It is probable that, if there is a lack of competitive forage and during the periods of high demand for protein, honey bee colonies and other bees use all possible sources of pollen (and starch for energy purposes). And during excess of food they sellect flowers. Still, the point is how they sellect? By colour, by amount, by scent, by access posibility (flower morphology), by ....??? And  what is the order of selection?  May be it is a simple question for behavioral physiologist, but not for 'botany people'

Interesting discussion. Bees will always select flowers even during excess of food, because optimal foraging translates into reproductive success. The quality of food is only one aspect amonst others, e.g. handling time to gather the food, distance from the hive, detectability of the food source (flowers). And we will probably end up with the honesty of floral signals: Do distinct floral signals honestly indicate the amount of reward? Floral colour changes obviously do. The amont of floral resoures which are earned best only once in the morning like the resin of Dalechampia flowers are honestly indicated by bract size. But what about "normal" flowers in which nectar reward can be frequently refilled: Is there any signal that may honestly indicate the standing crop, probably not? I there are differences amonst flowers in terms of refilling rate (fast - slow) and amount (high - low)? Does it make sense to indicate these features by means of an honest signal if the flowers are visited frequently by different visitors and almost no flower is completely filled to the max?

Please refer to:

Heiduk et al. (2010). Scent chemistry and pollinator attraction in the receptive trap flowers of Ceropegia. South African Journal of Botany 76(4):762-769.