In the last 70 years, the agricultural area devoted to pollinator-dependent crops has increased monotonically (Aizen et al., 2019). Animal pollination, mostly bee pollination, directly affects the yield of 87 of 115 leading single crops (Klein et al., 2007). Given the central place that pollination services have achieved in agriculture, it is necessary to improve the efficiency of pollinators, in particular of those managed by humans, like the honey bee. The challenge of improving pollination services depends on several factors, including knowledge about plant biology and pollination requisites, landscape features, environmental conditions, as well as the pollinator needs (McGregor, 1976; Free, 1993; Delaplane and Mayer, 2000; Abrol, 2012). In particular, improving pollination by managed honey bees requires monitoring the colonies introduced into the crop to assess levels of foraging activity and the resources collected before, during and after the blooming period. This knowledge allows the design of a pollinator management strategy to define the number, placement, and timing of colony introduction to obtain high yields. In addition, healthy and populous colonies are essential to ensure the success of pollination service (McGregor, 1976; Free, 1993; Delaplane and Mayer, 2000; Abrol, 2012).

The crop requirements and the management of its pollinators are covered within the topic known as “managed pollination” or “directed pollination” (Delaplane and Mayer, 2000; Vásquez Romero et al., 2011; Rosa et al., 2018). However, this area does not consider aspects of the individual and social behavior of honey bees, the most commonly managed pollinator worldwide, which directly impacts the pollination services provided. For instance, honey bees exhibit behaviors that advertise and recruit nestmates to the most profitable food sources. In this sense, honey bees have the ability to communicate spatial information about profitable sites through the waggle dance (a figure-of-eight maneuver on the vertical wax combs), and to transfer food-related information, such as scents or tastes, through mouth-to-mouth trophallactic food exchanges among nestmates (von Frisch, 1967; Farina et al., 2005). So far, honey bee plastic behavioral responses to new conditions required by crop pollination management, either by moving hives between environments that offer different floral availability, and/or after the sudden onset of a massive and dominant blooming, are seldom considered for pollination services. Furthermore, honey bees’ orientation and navigation abilities, as well as their capacity to learn floral-related cues, were often neglected. Within the behavioral sciences, these aspects are covered by cognitive ecology (Dukas, 1998), which considers how animals obtain and process information from their environments, and how they relate and use such information to make decisions according to their perception and learning abilities (Healy and Braithwaite, 2000).

Honey bees can visit a wide range of flower types as long as the resources offered are profitable (Visscher and Seeley, 1982; Steffan-Dewenter and Kuhn, 2003). Regardless of their generalist foraging strategy, honey bees exhibit fidelity to a single plant species within the same foraging bout (von Frisch, 1967). Such behavior is known as flower constancy (Free, 1963; von Frisch, 1967) and it implies that experiences with flowers that offer sufficient reward (pollen and/or nectar) encourage bees to keep collecting on the same floral type (Menzel and Erber, 1978; Chittka et al., 1999). It may vary according to the quantity and quality between the food sources (Wells and Wells, 1986). Thus, floral constancy together with the ability to communicate food-related information (location, profitability and chemosensory cues) within the nest (von Frisch, 1967; Farina et al., 2005), make the honey bee an efficient pollinator throughout a broad spectrum of agricultural settings (McGregor, 1976; Free, 1993).

The first attempts to improve food production in agricultural landscapes considering the plastic behavior of the honey bee were reported in the famine time before and during the World War II by different research groups from Germany and the ex-Soviet Union. In that time, different procedures to direct pollinators to target crops were based on the seminal study of von Frisch (1923), showing that recruited forager bees were prone to visit flowers of the same species previously exploited by scouting colony mates. von Frisch observed that the efficiency of the recruitment to a target feeding site depended on the distance to the hive and on the presence of floral odors, which were either emitted by parts of flowers attached to a feeder or diluted in the food. von Frisch noted that the offering of scented food enabled a better recruitment, likely because the chemical properties of the odors were better maintained until the liquid food was shared via trophallaxis with the colony mates. Pioneering practices to promote bee responses to target crops were then based on soaking fragrant flowers in sugar water, which produced a scented solution expected to bias foraging towards the target flowers (Smaragdova, 1933; Gubin, 1936; Gubin, 1938; von Frisch, 1943; von Frisch, 1947). The use of in-hive scented-food stimulation showed increases both in the number of bees that visited the crop (Gubin, 1938; Sorokin, 1938; Komarow, 1939) and in seed yields (Sorokin, 1938; von Frisch, 1947). von Frisch also tested the offering of scented sugar solution outside the hive (i.e., in the crop surroundings; von Frisch, 1943; von Frisch, 1947) and proved that this procedure was effective in increasing both the colony activity level, and the amount of honey produced (von Frisch, 1943), likely as it promotes the display of dances. However, a study of Free (1958) in apple and red clover crops using either the offering of scented sugar solution outside or inside the hive, or the combination of both, showed no evidence of increases in crop yields. Later, Free (1969) tested the extent to which the odor of nectar stored in combs affected foraging preferences in a double-choice test. Although the results were highly variable, Free was able to detect a brief biased response to the odor present in the honeycomb.

Despite their relative success in guiding bees to target crops, procedures that soak fragrant flowers in sugar water have several disadvantages, such as the poor stability of the odor extracted from the flowers and the cost involved in obtaining large quantities of flowers to achieve a stimulus sufficiently intense to modify bee responses. Furthermore, the cutting and crushing of flowers for the syrup preparation may promote the release of unwanted volatiles, related to tissue damage or wilting, being a strong source of variation among results of pioneering studies. For this reason, it is relevant to integrate aspects related to floral odors and honey bee social behavior as part of a targeted pollination strategy. With this in mind, the objective of this review is to summarize some pertinent elements related to individual and social honey bee learning of floral scents that affect foraging responses, which are potentially applicable for guiding bees to target crops to enhance pollination services. Floral volatiles of specific crops, honey bee odor perception, social foraging, and the procedures in the field will also be discussed as necessary components within the targeted pollination framework ( Figure 1 ).

Floral bouquets are complex mixtures of volatile organic compounds (VOCs) which are directly involved in plant-pollinator interactions (Knudsen et al., 1993; Raguso, 2008; Pichersky and Dudareva, 2020). Volatile emissions can be altered by several factors, such as cultivar, time of day and pollination status (Rodriguez-Saona et al., 2011; Twidle et al., 2017). Depending on the identity and concentration of the VOCs emitted by the different plant species, diverse specific pollinator groups are attracted (Dobson, 2006). In particular, the olfactory cues that honey bees use to perceive specific flowers has been investigated for different crops, such as oilseed rape (Wadhams et al., 1994), kiwifruit (Twidle et al., 2015), pear (Su et al., 2022) and other Brassicaceae species (Kobayashi et al., 2012), among others. In such studies, honey bee odor detection was assessed by means of electro-antennography (EAG) where the antennal response towards the different floral bouquets is measured. It is well known that, although plants emit large amounts of VOCs, honey bees detect a small subset of these compounds or key odorants (Reinhard et al., 2010; Mas et al., 2020). Moreover, not all odors detected by the peripheral olfactory system are behaviorally meaningful, and most must be learned before they can influence behavior (Riffell et al., 2009b; Menzel, 2012).

Many odors, which initially are neutral to honey bee foragers, may become good predictors of food sources after being learned (von Frisch, 1967; Lindauer, 1970; Gould, 1984; Menzel, 2012). Learning allows individuals to flexibly respond to a changing environment (Menzel, 1999), with varying availability of food sources during the season (Núñez, 1977; Vogel, 1983), being extremely important in species with generalist habits. In this way, honey bees as well other pollinators are able to associate floral cues, such as odors and colors, with the rewards (nectar, pollen) that the source provides (Gould, 1984; Chittka and Thomson, 2001). If bees repeat cue-reward experiences, these associations turn into memories that influence foraging behaviors, by biasing flight orientation (Chaffiol et al., 2005; Nery et al., 2021), landing (Arenas et al., 2007; Arenas et al., 2008), and/or extension of the proboscis (Grüter et al., 2006; Arenas and Farina, 2012). The latter is an innate reflex response that occurs when a bee’s antennae contact the nectar of a flower, leading to an immediate ingestion of the food. In the laboratory, the proboscis extension response (PER) can be evoked by touching the antennae of restrained bees with an enough concentrated sucrose solution (Kuwabara, 1957; Takeda, 1961). Moreover, bees can be trained to associate an odor with a sucrose reward, by means of an olfactory conditioning protocol (Takeda, 1961; Bitterman et al., 1983). Prior to conditioning, bees do not usually respond to the conditioned stimulus (CS), but after successive paired presentations between the odor and a sucrose reward, the previously neutral stimulus now takes control over the proboscis reflex.

Within the PER paradigm, it is possible to train bees to learn that an odor predicts an oncoming reward in the absence of other alternative stimuli, in the so-called absolute conditioning (Giurfa, 2007; Giurfa and Sandoz, 2012). This protocol allows us to test the extent to which the conditioned response can be generalized to different but equivalent stimuli. The phenomenon of generalization is widespread among animal kingdom (Shepard, 1987; Ghirlanda and Enquist, 2003), and it is essential for the foraging behavior of the bees since it enables foragers to respond to different floral odors if they are perceived as similar (Pham-Delegue et al., 1989; Guerrieri et al., 2005). Alternatively, bees trained in a differential conditioning learn not only the characteristics of a reinforced stimulus (rewarded conditioned stimulus, henceforth: CS+), but also those of a nonreinforced one (non-rewarded conditioned stimulus, henceforth: CS-) (Bitterman et al., 1983; Giurfa, 2007; Giurfa and Sandoz, 2012). This protocol allows us to evaluate bees’ abilities to discriminate between both conditioned stimuli. Combined, the two types of conditionings enable investigating how bees learn, generalize and discriminate odorant mixtures.

Previous studies about insect behavior demonstrated that plant-pollinator interactions can be mediated by a few key odorants (Riffell et al., 2009a; Riffell et al., 2009b; Reinhard et al., 2010; Mas et al., 2020). Using the moth Manduca sexta as a study model, it was observed that food source attraction and innate foraging behavior could be elicited by a few of the volatile compounds that conform the natural flower bouquet (Riffell et al., 2009a; Riffell et al., 2009b). The evidence that a few key volatiles are sufficient to account for a complex odor mixture is not limited to innate behaviors but extends to learned responses as well. In honey bees, response to mixtures composed of a few selected key odorants (some of only 3 pure compounds) were sufficient to elicit levels of PER comparable to those evoked by the complete olfactory mixtures of 14 odorants to which the bees were initially conditioned (Reinhard et al., 2010). The key odorant processing of floral scents may be adaptive to maintain stimulus identity in a constantly changing environment while it gives us the possibility for using simple mixtures to mimic the complex floral scent of a species of interest to manipulate odor-mediated responses of pollinators. In fact, it has been recently shown that the conditioning of synthetic mixtures with 3 or 4 constituents could be enough to successfully generalize the natural floral scent of agriculturally important species, such as sunflower, pear, apple, and almond (see Figure 2A as example), which in turn resulted in a bias of the bees’ foraging behavior towards the target crop (Farina et al., 2020; Farina et al., 2022; Farina et al., 2023).

Apart from evaluating the degree of generalization of specific odor mixtures designed to mimic the scent of crop flowers, Farina and coworkers (2020; 2022; 2023) assessed to what extent bees could discriminate them from the respective natural floral scents. Differential PER conditionings using the mimic odor and the natural floral scents both as CS+ and CS- revealed that bees could only discriminate between the stimuli when the natural blend was presented as CS+ and each mimic (either for the sunflower, pear, or apple flower) as CS-. Interestingly, bees failed to distinguish between stimuli if the mimics acted as CS+ and the natural blend as CS-, indicating that discrimination was not symmetric (see Figures 2B, C as example). Such asymmetry between a small subset of key odors that make up a behaviorally effective mixture and the natural floral bouquet denotes the complexity of insect olfactory perception (Sandoz et al., 2001) and suggests that a mimic odor could be more effective in biasing foraging behavior if learned while bees remain naive to crop flower´s olfactory cues than in experienced individuals.

Considering that memories decay in time (Menzel, 1999), some studies focused on the effect of the addition of nonsugar nectar compounds on honey bee olfactory associative learning with the aim to establish more stable long-term memories (Wright et al., 2013; Marchi et al., 2021). On the one hand, the alkaloid caffeine triggers long-term memory in honey bee (Wright et al., 2013), meanwhile the essential amino acid arginine participates in the synthesis of nitric oxide, and therefore promotes protein synthesis during long-term memory formation (Müller, 1996; Müller, 1997). A recent related study showed a positive effect of the combination of caffeine and arginine in the reward, increasing bees’ learning performance and long-term memory formation (Marchi et al., 2021). Thus, the joint administration of nonsugar nectar compounds with synthetic mimic odors could further enhance the persistence of olfactory memories and therefore, the efficacy of a procedure that aims to modify bee´s preferences based on experience.

A honey bee colony can rapidly adjust its foraging behavior and guide its workforce toward the most rewarding flowers in the surrounding environment (Seeley, 1995). Within the hive, social interactions among nestmates facilitate the propagation of food-related information, allowing not only experienced foragers to access information about other available sources, but also new recruits to locate profitable foraging sites via the waggle dance (von Frisch, 1967). There is a consensus that olfactory cues of the discovered resource play an important role in orientation at short distances (von Frisch, 1923; Sherman and Visscher, 2002). On the other hand, further studies support the idea that olfactory cues alone are not sufficient to recruit nestmates (Gould, 1984; Riley et al., 2005) and that orientation of foragers fails if olfactory stimuli from the food source to which they have been trained are relocated beyond 200 meters (Menzel and Greggers, 2013). However, odorants can assist recruits to reach the target if they are learned in the colony and within a recruiting context (Farina et al., 2005; Díaz et al., 2007).

Beside the transmission of spatial information, the dance increases the attention and activity of bees in the vicinity, attracting them to the dancer (von Frisch, 1967; Grüter and Farina, 2009; Balbuena et al., 2012a; Ai and Farina, 2023). Then, more, and highly motivated bees around the dancer can learn the floral odor molecules attached to its body (Moauro et al., 2018). Dancing bees often briefly interrupt the dance and offer food samples to surrounding bees (von Frisch, 1967; Díaz et al., 2007). These oral interactions (i.e. trophalaxis) can be very brief, but just long enough to act as a reward in olfactory learning (Díaz et al., 2007; Farina et al., 2007; Farina and Grüter 2009; Grüter and Farina, 2009). For scented nectars, trophallactic interactions enable the establishment of memories from odors diluted in the food that is being shared, an effective mechanism when scouts forage from distant sources while the odors attached to their body fade during the trip back to the hive. For pollen foragers, cues associated with pollen loads carried on the hind legs of dancers may also be perceived and learned by other foragers (Díaz et al., 2007; Nery et al., 2020) giving selectivity to recruitment (Arenas et al., 2021).

Memorization of olfactory cues within the nest, albeit outside the dancing context, could also assist recruits locate the feeding site (Balbuena et al., 2012b). Olfactory cues could also be learned from scented nectars that are unloaded to the food processor bees (Grüter et al., 2006; Grüter et al., 2009). The food odors learned inside the nest can be retained by colony mates for up to 10–11 days suggesting that olfactory experiences occurring within the colony can propagate to many individuals (Grüter et al., 2009). Moreover, circulation of scented sugar solution biases foraging preferences towards the learned odor, a response that is extended until four days after removing the scented-food stores and the combs where the syrup could have been stored (Arenas et al., 2007; Arenas et al., 2008). It is not trivial to mention that when the odor is not offered in the food but presented as a volatile that aromatizes the nest environment (Arenas et al., 2008), an avoidance rather than an improvement of the landing response towards the exposed odor is observed. These results suggest that the presentation of odors, unpaired with the reward, triggers cognitive processes other than associative learning, which prevents the nectar foragers to visit sources scented with the exposed odor. In summary, although other sensory modalities (e.g. visual) may be much more effective for long-distance flights during searching resources (Chittka and Menzel, 1992; Dyer et al., 2011; Menzel and Greggers, 2013), the use of floral odors via olfactory memories are crucial in the search for food sources when combined with other social interactions occurring in the nest.

Given that olfactory information transfer can also occur when odors are directly provided inside the nest (Arenas et al., 2007; Arenas et al., 2008), the offering of scented food can be used as a standardized procedure to establish specific long-term memories among foragers being part of a targeted pollination strategy. A common practice of beekeepers is feeding colonies with sugar syrup at certain times of the year, for instance during dearth periods of nectar (Geslin et al., 2017a; Sammataro and de Guzman, 2018; FAO et al., 2021). Furthermore, feeding colonies with syrup in the fall can ensure survival through winter and the provision of sugar syrup inside the hive stimulates brood rearing and thereby promotes foraging for pollen (Goodwin, 1997; Sammataro and de Guzman, 2018). Sugar syrup can be offered by means of in-hive feeders of different types, such as division board or top hive feeders. Such supplemental feeding practice is suitable for olfactory conditioning of colonies providing pollination services (Farina et al., 2020; Estravis-Barcala et al., 2021; Farina et al., 2022; Farina et al., 2023). Scented food can be obtained by diluting a small volume of the mimic odor (50 µL) per liter of sucrose solution (50% weight/weight, henceforth: w/w). Scented syrup can be offered using in-hive feeders or even poured over the top of the central frames of the hives (for 1,000-1,500 mL or 500 mL of scented solution, respectively).

Some important aspects to be considered in honey bee management for pollination services are colony size (Goodwin, 1997; Geslin et al., 2017a; Ovinge and Hoover, 2018; Chabert et al., 2021) and the timing of colony introduction into the plots (Free, 1959; Free et al., 1960; Al-Tikrity et al., 1972; Moeller, 1973; Sammataro and de Guzman, 2018). If colonies are introduced into the agricultural setting long before blooming, bees could forage on other attractive non-target flowers and may ignore the crop when it blooms. On the other hand, if colonies are settled when the crop is already in full bloom, they may not be able to learn the cues related to the crop flowers before blooming ends and may not have enough time to learn the landmarks needed to orient themselves. The timing of stimulation is critical as well. Feeding should be done at the beginning of the blooming period to guarantee bees an early access to relevant olfactory information which will assist them in finding the target flowers in a novel environment. It is advisable to perform the stimulation of colonies when the target crop is 10-40% in bloom. A single stimulation event should suffice to guide bees to the target crop. But it should be considered that the number of events may vary with the specific requirements of the crops and the weather conditions.

The targeted pollination strategy has the advantage of being specific to the crop, and usually requires only one application (Farina et al., 2020; Farina et al., 2022; Farina et al., 2023). This method facilitates the propagation of food related information among nestmates, which can be retrieved several days later. These studies demonstrate that specific olfactory memories established within the honey bee colony result in faster foraging that increases crop production, showing the advantages of a targeted pollination approach to enhance pollination services in commercial crops (see Supplementary Table S1 ).

It should be mentioned that there is a study testing the method of osmoguiding bees with a maceration and cooking of crop flowers, which failed to promote visits to the target pollen (Higuera-Higuera et al., 2023). So far, these results are inconclusive, as some of the assays need more controls to be confirmed.

In the last decades, several management methods were developed in an attempt to improve honey bee pollination of crops with ambiguous results (Goodwin, 1997; Delaplane and Mayer, 2000). Recently, the use of mimic odors based on crop floral volatiles has proven to be successful in guiding honey bees toward a target crop, which in turn positively affected foraging activity in systems highly dependent on pollinators, such as sunflower for hybrid seed production, apple, pear and almond, and consequently, increased yields (Farina et al., 2020; Farina et al., 2022; Farina et al., 2023). Although the mentioned species are mass-flowering crops, offering plentiful floral resources for bees, they differ in the type of plantation. While sunflower and almond are usually grown on large-scale monoculture fields and orchards (Farina et al., 2020; Estravis-Barcala et al., 2021; Farina et al., 2023), apple and pear trees are cultivated at a much smaller scale (< 10 ha) and sometime coexist within the same orchard (Díaz et al., 2013; Quinet et al., 2016).

The growing global demand for pollination services (Aizen et al., 2019) leads to propose new strategies in honey bee management to improve its efficiency in agroecosystems. The implementation of a targeted pollination strategy mediated by honey bee plastic responses integrates aspects related to floral odors and honey bee social behavior, including communication processes. Within this framework, the results so far obtained suggest that conditioning bees to simple synthetic odorant mixtures which mimic specific flowers could enable the establishment of in-hive odor memories that bias bees to the target crop and potentially increase yields (Farina et al., 2020; Farina et al., 2022; Farina et al., 2023). From the growers’ perspective, this method might decrease the honey bee stocking rate by increasing the pollination activity of the honey bee colonies and therefore to save on input costs. While at the same it can help to decrease the detrimental effects of managing too many honey bees at the same location on wild flora and entomofauna (Geslin et al., 2017b; Morales et al., 2017; Russo et al., 2021). Nevertheless, it is worth remarking that there are knowledge gaps to be further investigated. The use of volatile mixtures as odor mimic could be challenging for crops that involve different varieties and this will require a thorough understanding of cultivar-specific floral bouquets (Rodriguez-Saona et al., 2011; Twidle et al., 2017). Additionally, specific crops might present certain characteristics detrimental to pollination by honey bees (i.e., brief blooming period, restrictive floral morphology, lack of nectar as reward). Also, future works are necessary to determine how long the effect of the treatment of feeding colonies with sugar syrup scented with mimic odors lasts. At the same time, it remains to be assessed the extent to which this procedure can be implemented with alternative managed bees for those crops where honey bees are less efficient pollinators. The honey bee is not the most efficient pollinator in many cases (Ne'eman et al., 2010) due to quite limited single visit pollen depositions (Földesi et al., 2021; Page et al., 2021), because they are not especially effective to transfer cross-pollen on cultivars requiring cross-pollination, especially when wild entomofauna is absent (Garibaldi et al., 2013), or they forage on a large area around their nest, resulting in a high probability to be diverted to other competing bloom (Jay, 1986; Quinet et al., 2016; Osterman et al., 2021a). This is of particular interest since many native bees (e.g., bumble bees and solitary bees) are currently reared for agricultural purposes (Osterman et al., 2021b). Lastly, although this procedure has great potential for positive impacts on food industry, research on a proper packaging to maintain the chemical stability of the mixture will also be needed to determine its economic viability. The economic impact of an efficient and sustainable entomophilous pollination procedure for the most high-market valuable crops could improve yields in quantitative and qualitative terms in a global context of increasing demand of pollination services.

All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.