Searching for reward motivates and drives behaviour. In honey bees Apis mellifera, specialized pollen foragers are attracted to and learn odours with pollen. However, the role of pollen as a reward remains poorly understood. Unlike nectar, pollen is not ingested during collection. We hypothesized that pollen (but not nectar) foragers could learn pollen by sole antennal or tarsal stimulation. Then, we tested how pairing of pollen (either hand- or bee-collected) and a neutral odour during a pre-conditioning affects performance of both pollen and nectar foragers during the classical conditioning of the proboscis extension response. Secondly, we tested whether nectar and pollen foragers perceive the simultaneous presentation of pollen (on the tarsi) and sugar (on the antennae) as a better reinforcement than sucrose alone. Finally, we searched for differences in learning of the pollen and nectar foragers when they were prevented from ingesting the reward during the conditioning. Differences in pollen-reinforced learning correlate with division of labour between pollen and nectar foragers. Results show that pollen foragers performed better than nectar foragers during the conditioning phase after being pre-conditioned with pollen. Pollen foragers also performed better than nectar foragers in both the acquisition and extinction phases of the conditioning, when reinforced with the dual reward. Consistently, pollen foragers showed improved abilities to learn cues reinforced without sugar ingestion. We discussed that differences in how pollen and nectar foragers respond to a cue associated with pollen greatly contribute to the physiological mechanism that underlies foraging specialization in the honeybee.

Animals learn to respond to stimuli that are reliable in predicting the reward. How they perceive the reward and learn stimuli is thought to be critical in terms of accounting for division of labour in social groups (Dukas, 1998; Couzin, 2009; Couzin, et al., 2011; Sih and Del Giudice, 2012). Division of labour finds its maximum expression in social insects, where workers specialize on subsets of the tasks performed by the colony (Wilson, 1971, 1987; Oster and Wilson, 1978). Task division is usually associated with morphological and/or physiological differences besides changes in behaviour throughout the lifespan of individuals (Robinson, 1992; Seeley, 2009; Hölldobler and Wilson, 2008; Duarte et al., 2011; Hammel et al., 2016). Evidence from honey bees, ants, social wasps and stingless bees suggests that the probabilities of engaging in a specific task are linked to individual differences in response thresholds to task-related stimuli (Robinson and Page, 1989; Beshers and Fewell, 2001; Perez et al., 2013; Mattiacci, 2019; Balbuena and Farina, 2020). Individuals with a relatively low threshold for a given task tend to respond to lower stimulus intensities and, consequently, to learn better if the stimulus acts as reinforcement (e.g. food) (Scheiner et al., 2005; Perez et al., 2013). On the contrary, individuals with higher thresholds learn worse about reinforcements to which they are less sensitive.

In the honey bee, collection of carbohydrates and protein is achieved by individuals that specialize in either nectar or pollen foraging (Winston, 1987). Foraging task specialization between nectar and pollen foragers has been correlated with differences in response thresholds to sucrose stimulation (Page et al., 2006), quantified as the lowest concentration of sugar solution that elicits the proboscis extension reflex (PER; Page and Fondrk, 1998). Even when it sounds counterintuitive, nectar foragers exhibit higher thresholds (i.e. lower responsiveness) than pollen foragers. This feature might be adaptive, as nectar foragers selecting the most concentrated sources would provide the highest energy gain to the colony. However, differences in sucrose responsiveness do not fully explain why foragers with lower thresholds for sugar are more prone to collect pollen.

Even when many studies treat pollen and nectar rewards as equivalents, very little is known about the function of pollen as reinforcement of learning in honey bee foragers (Grüter et al., 2008; Arenas and Farina, 2012, 2014; Nicholls and Hempel de Ibarra, 2013). Whilst foraging nectar, bees associate the surrounding cues with the energetic reward that the source provides. However, and because foragers rarely consume pollen at the food source and transport it in external structures located on the third pair of legs (i.e. corbiculae), it is less likely for pollen to act via ingestive reward pathways (i.e. post-ingestion). Nevertheless, ingestion might not be critical for pollen reinforcement, and learning might also occur pre-ingestion (Sandoz et al., 2002), when the antennae, mouthparts or tarsi repeatedly come in touch with pollen grains during gathering, thus reinforcing associative learning through similar gustatory pathways as sucrose. Because pollen foragers exhibit significantly higher sensitivity to gustatory stimuli than nectar foragers (e.g. for sucrose; Page and Fondrk, 1998; Pankiw and Page, 2000), they might be able to perceive chemosensory cues of pollen (Scheiner et al., 2004) as excitatory stimuli that mediate learning and memory.

In experiments that resemble natural foraging conditions, pollen foragers were able to associate a neutral odour with pollen, and these memories guided the bees towards the source that smelt like the learned odour (Arenas and Farina, 2012). Unfortunately, these approaches do not allow us to precisely identify the nature of the reinforcing components of pollen, or to control the sensory organs and timing of stimulus and reward presentations. Another no less important limitation is that no comparison can be made between foragers specialized in either pollen or nectar collection, which is essential for understanding the physiological differences underlying foraging division of labour. To address some of these points, previous studies were conducted under the PER paradigm (Grüter et al., 2008; Arenas and Farina, 2012; Nicholls and Hempel de Ibarra, 2013), which was extensively used to study learning and memory in the honey bee with sugar as a reward. It is based on the conditioning of a reflex, primarily triggered by a sugar solution applied on tarsi or antennae (Takeda, 1961; Bitterman et al., 1983). The contingency created between the sucrose (unconditioned stimulus, US) and the odour (conditioned stimulus, CS) enables the bee to predict the oncoming reward by extending the proboscis as a sole response to the CS.

Honey bees, especially pollen foragers, could also extend their proboscis when stimulated with hand- and bee-collected pollen (Scheiner et al., 2004; Arenas and Farina, 2012; Grüter et al., 2008); however, contradictory results have been obtained when trying to condition bees with pollen. Using bee-collected pollen, Grüter and colleagues (2008) reported that approximately 70% of the pollen foragers (i.e. bees that returned to the hive with pollen loads) were able to associate an odour with a hydrated bee-collected pollen reward. Conditioned response was also achieved using drunken stick (Ceiba speciosa) hand-collected pollen but failed with dandelion (Arenas and Farina, 2012). Experiments that did not discriminate between pollen and nectar foragers reported that despite steady levels of extension triggered by dry bee-collected pollen, bees did not learn to associate both stimuli (Nicholls and Hempel de Ibarra, 2013). Given these results, a variety of factors may challenge what we know about the role of pollen as a reinforcement during olfactory learning, namely: where pollen came from (i.e. either hand- or bee-collected), where and how it is presented (i.e. in the antennae or in the tarsi, with or without reward ingestion) and to whom it is presented (i.e. either nectar or pollen foragers).

Furthermore, it has been observed that the odour–pollen association learnt in a pollen source could bias orientation in a ‘Y’ maze (operant context) but could not be transferred to the PER paradigm (Arenas and Farina, 2012). So far, and even though bees use the proboscis and the mandibles for pollen collection, whether information learned with pollen could be evoked through the PER is open to debate. Nonetheless, pairing of odour and pollen might occur irrespective of behavioural responses to the odour and the reward, thus olfactory memories learnt with pollen could still be testable in the PER paradigm if pollen and sugar share (fully or partially) the same reinforcement processing pathways.

Here, we aimed to study the reinforcing properties of pollen, to answer whether it can excite the bees in a way that enables or enhances learning and memory in the classical context of the PER. In addition, we tested the hypothesis that differences in pollen- reinforced learning correlate with division of labour between pollen and nectar foragers. In a first experiment, we addressed the role of pollen as an excitatory stimulus during olfactory learning. To this end, pollen was paired with an odour under the PER paradigm, in what we called a pre-conditioning phase. Both hand-collected and bee-collected pollen were used as reinforcement, either presented in the antennae or tarsi of pollen and nectar foragers. Considering that pollen might not always trigger PER, yet may be effective as reinforcement for the bees, putative olfactory memories established during the pre-conditioning were evaluated in a classical PER conditioning, in which the odour previously paired with pollen was now conditioned with a sucrose reward. In a second experiment, we evaluated the interaction between pollen and sugar rewards on PER conditioning. Then, nectar and pollen foragers were olfactory conditioned using a sucrose solution (presented on the antennae) plus hand-collected pollen (presented on the tarsi). We predicted that the simultaneous presentation of pollen and nectar would lead to higher acquisition and/or lower extinction rates than sucrose alone, owing to an additive or synergic effect of the dual sensory input. Finally, we investigated the bee's capacity to form associative memories regardless of sugar reward ingestion. Because pollen foragers do not ingest pollen at the food sources, we predicted that they would perform better than nectar foragers during olfactory learning when they are prevented from assessing the resource through ingestion.

Study site and animals

Experiments were carried out during the summer seasons of 2016–2017 and 2017–2018 in the experimental field of the Faculty of Exact and Natural Sciences of the University of Buenos Aires (34°32′S, 58°26′W). Free-flying worker bees of unknown age that belonged to colonies of European honeybees Apis mellifera Linnaeus 1758 were used for the experiments. Pollen foragers were captured during the mornings of the experimental days at the hive entrance, while returning with pollen loads. Nectar foragers were captured when they arrived at an ad libitum feeder with 10% w/w sucrose–water solution. We chose a 10% w/w to minimize differences in gustatory sensitivity between the sub-castes. Honeybees were chilled in a freezer, and pollen loads were removed from hind legs. Afterwards, bees were individually restrained in harnesses in such a way that they could freely move their antennae, mouthparts, and forelegs or hind legs. All bees were fed with sucrose–water solution until satiation and were kept in an incubator (30°C, 60% relative humidity and darkness) for 2 h. All experiments complied with the animal care guidelines of the National Institutes of Health (1985) and the current laws of Argentina.

PER conditioning protocol

The conditioning of the PER is based on the contingency created between the reward (unconditioned stimulus, US) and the odour (conditioned stimulus, CS) that enables the bees to extend their proboscis as a sole response to the odour, predicting the oncoming reward (Kuwabara, 1957; Bitterman et al., 1983). Linalool (LIO) was used as the CS, a neutral scent that is present in flowering plants (Raguso, 2008) and has been widely used as CS in olfactory conditioning experiments. The CS was delivered by means of an automated device that sends a continuous clean air flow (50 ml s−1) to the bee's head. A controlled valve deflected part of the air flow (6.25 ml s−1) through a 0.5 ml syringe that contained a small piece of filter paper saturated with 4 μl of LIO, which injected the flow again into the main clean air flow. Each trial lasted 55 s. The valve was programmed so that it released clean air during the first 20 s, followed by the odour for 6 s and clean air again for the remaining 29 s. We measured whether the bees exhibited the proboscis extension or not during the first 3 s of the odour presentation. Only bees that did not show response to a mechanical stimulus (airflow) were tested. Sucrose–water solution was used as the US, and the concentration of the solution depended on the nature of each experiment.

Experiment 1. Classical conditioning after pre-conditioning with pollen

In this experiment, we aimed to test whether foragers are able to learn an odour using pollen as the rewarded stimulus in a classical olfactory context and to evaluate whether there are differences in learning between pollen and nectar foragers. Considering that pollen might not always elicit a response in the honeybees, yet be effective at reinforcing learning, we first performed a pre-conditioning with pollen as reinforcement and then we evaluated whether such a pre-conditioning affected the subsequent learning performance in a classical conditioning with sucrose as a reward (Fig. 1A). Nectar and pollen foragers were subjected to a pre-conditioning phase that consisted of four trials, where we paired the presentation of the CS (Fig. 1B) to one of three different stimulations: (1) hand-collected kiwi pollen on the antennae, (2) fresh honey bee-collected pollen (unknown flower sources) on the antennae or (3) hand-collected kiwi pollen on the tarsus of the left hind leg. To prevent differences in performance being due to sensitization after pollen presentation, groups of pollen foragers were also subjected to an unpaired pre-conditioning (Fig. 1C), where pollen was presented 14 s before the CS was delivered. Memories that could have been formed during the pre-conditioning with pollen were then evaluated in a classical PER conditioning protocol of three trials, where we paired the odour with a sucrose–water solution (50%) reward (Fig. 1D).

Fig. 1.

Experiment 1: classical conditioning after pre-conditioning with pollen. Procedure to examine associative learning with pollen in honey bees. (A) General procedure for testing the bees. (B) Detail of the paired pre-conditioning with pollen. (C) Detail of the unpaired pre-conditioning with pollen. (D) Detail of the paired classical conditioning with sucrose solution after the pre-conditioning phase.

Fig. 1.

Experiment 1: classical conditioning after pre-conditioning with pollen. Procedure to examine associative learning with pollen in honey bees. (A) General procedure for testing the bees. (B) Detail of the paired pre-conditioning with pollen. (C) Detail of the unpaired pre-conditioning with pollen. (D) Detail of the paired classical conditioning with sucrose solution after the pre-conditioning phase.

Experiment 2. Interaction between pollen and sugar reward on PER conditioning

To evaluate pollen and sugar solution interactions on learning and memory, we performed the classical olfactory conditioning of the PER using a double stimulation of pollen and sucrose solution as reward. Bees were olfactory conditioned during five training events under three different procedures: (1) paired pollen procedure, (2) unpaired pollen procedure and (3) non-pollen procedure (Fig. 2). In the first procedure (paired pollen procedure) the presentation of the odour was paired with both hand-collected kiwi pollen (offered on the first tarsi) and sucrose–water solution 30% (presented on the antennae; Fig. 2B). On the second procedure (unpaired pollen procedure), the presentations of the odour and the sucrose solution were paired, but unpaired with pollen, which was presented 6 s after the end of the odour and sugar presentations (Fig. 2C). In the last procedure (non-pollen procedure) there was no pollen involved. Here, the odour presentation was paired with the sucrose solution (Fig. 2D). Memories formed during the PER conditioning were then evaluated for their extinction along four trials that consisted of the presentation of the odour alone (Fig. 2E).

Fig. 2.

Experiment 2: interaction between pollen and sugar reward on PER conditioning. Procedure to examine the interaction between pollen and sugar reward on PER conditioning. (A) General procedure for testing the bees. (B) Detail of paired pollen procedure. (C) Detail of unpaired pollen procedure. (D) Detail of the non-pollen procedure. (E) Detail of the extinction procedure.

Fig. 2.

Experiment 2: interaction between pollen and sugar reward on PER conditioning. Procedure to examine the interaction between pollen and sugar reward on PER conditioning. (A) General procedure for testing the bees. (B) Detail of paired pollen procedure. (C) Detail of unpaired pollen procedure. (D) Detail of the non-pollen procedure. (E) Detail of the extinction procedure.

Experiment 3. Classic olfactory conditioning without reward ingestion

Under the hypothesis that pollen foragers can learn better than nectar foragers with a pre-ingestive reward, we tested for any difference in performance during a classical olfactory conditioning where sucrose was only presented on their antennae. Following the standardized PER protocol for olfactory conditioning described above, both pollen and nectar foragers were submitted to four acquisition trials, in which the odour delivery was paired with the presentation of a 50% w/w sucrose solution on the antennae. It is worth mentioning that any contact with the proboscis was avoided. In addition, memory retention was tested 1 h after finishing the acquisition phase in a non-rewarded single PER trial.

Data analysis

All statistical analyses were performed with R v.3.3.3 (R Development Core Team 2019) via RStudio (RStudio Inc. 2019). For experiments 1 and 2, the number of PERs to the conditioned odour elicited by each bee was summed and used as the response variable. In the pre-conditioning of experiment 1, it ranged from 0 to 4, and from 0 to 3 in the following conditioning. During conditioning of experiment 2, it ranged from 0 to 5, and from 0 to 4 during the extinction phase. For experiments 1 and 2, we used a generalized linear mixed model (GLMM) following a quasi-binomial error distribution. For experiment 3, we used a GLMM for a binomial distribution. For experiments 1 and 2 we used the ‘glm’ function of the ‘lme4’ package (Bates et al., 2015), and for experiment 3 we used the ‘glmer’ function of the ‘lme4’ package (Bates et al., 2015; Lenth, 2015). For the PER training we considered the type of forager (two-level factor corresponding to pollen and nectar forager) and trial (four-level factor corresponding to 1–4 trials) as fixed effects, and each individual as a random factor. Tukey’s test comparisons were used when needed.

Experiment 1. Classical conditioning after pre-conditioning with pollen

Stimulation of the antennae with hand-collected kiwi pollen

A very low percentage of the honey bees elicited the PER when stimulated with the kiwi pollen (PF_PP=18%; PF_UP=29%; NF=0%; where PF is pollen foragers, NF is nectar foragers, PP is paired procedure and UP is unpaired procedure); therefore, it was not possible to determine directly whether the pollen–odour association was established. Consequently, levels of conditioned response during pre-conditioning were low and did not change throughout the trials in any of the three groups (Fig. 3A; Table S1). During the following phase, conditioning with sugar, all individuals presented the unconditioned response to the sucrose. The percentages of the conditioned response increased through the trials in the three groups (Table S1), although with varying performances. The performance of the pollen foragers was better than that of the nectar foragers in the paired procedure [Pr(>|z|)=0.015]. On the contrary, no significant differences were either found between the PF_PP and PF in the unpaired training procedure (UP) [Pr(>|z|)=0.822] or between the PF_UP and NF_PP groups [Pr(>|z|)=0.051] (Fig. 3B).

Fig. 3.

Experiment 1: classical conditioning after pre-conditioning through stimulation of the antennae with hand-collected kiwi pollen. (A) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the four trials in which the conditioned odour was paired [circles for pollen foragers (PF_PP), squares for nectar foragers (NF_PP)] or unpaired (triangles for pollen foragers (PF_UP) with the pollen stimulation (pre-conditioning). (B) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the three trials in which the conditioned odour was paired with a sucrose reward (conditioning with sugar). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (***P<0.001).

Fig. 3.

Experiment 1: classical conditioning after pre-conditioning through stimulation of the antennae with hand-collected kiwi pollen. (A) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the four trials in which the conditioned odour was paired [circles for pollen foragers (PF_PP), squares for nectar foragers (NF_PP)] or unpaired (triangles for pollen foragers (PF_UP) with the pollen stimulation (pre-conditioning). (B) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the three trials in which the conditioned odour was paired with a sucrose reward (conditioning with sugar). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (***P<0.001).

Stimulation of the antennae with bee-collected pollen

In this experiment, bee-collected pollen was used as the reward during the pre-conditioning phase, and unlike stimulation of the antennae with hand-collected kiwi pollen, it enabled the conditioning of the PER among pollen foragers. The average unconditioned response to pollen was 68% for pollen foragers in both paired and unpaired procedures and 14% for nectar foragers. Fig. 4A shows that the levels of conditioned response for the PF_PP group was higher than the levels of response acquired by the other two groups [PF_PP–NF_PP: Pr(>|z|)<0.001; PF_PP–PF_UP: Pr(>|z|)<0.001]. During the conditioning with sugar phase (Fig. 4B), all groups showed some level of acquisition (Table S1); however, the performance of the PF_PP group was better than the other groups [PF_PP–NF_PP: Pr(>|z|)=0.002; PF_PP–PF_UP: Pr(>|z|)=0.001], indicating that pre-conditioning affected the performance of this group. Finally, no significant differences were found between the PF_UP and NF_PP groups.

Fig. 4.

Experiment 1: classical conditioning after pre-conditioning through stimulation of the antennae with bee-collected pollen. (A) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the four trials in which the conditioned odour was: paired [circles for pollen foragers (PF_PP), squares for nectar foragers (NF_PP)] or unpaired [triangles for pollen foragers (PF_UP)] with the pollen stimulation (pre-conditioning with pollen). (B) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the three trials in which the conditioned odour was paired with a sucrose reward for all groups (conditioning with sugar). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (***P<0.001; **P<0.01).

Fig. 4.

Experiment 1: classical conditioning after pre-conditioning through stimulation of the antennae with bee-collected pollen. (A) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the four trials in which the conditioned odour was: paired [circles for pollen foragers (PF_PP), squares for nectar foragers (NF_PP)] or unpaired [triangles for pollen foragers (PF_UP)] with the pollen stimulation (pre-conditioning with pollen). (B) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the three trials in which the conditioned odour was paired with a sucrose reward for all groups (conditioning with sugar). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (***P<0.001; **P<0.01).

Stimulation of the tarsi with hand-collected kiwi pollen

Here again, a very low percentage of foragers extended the proboscis when stimulated with kiwi pollen on the tarsi (PF_PP=5%; PF_UP=4%; NF=0%). Interestingly, those bees that did respond were pollen foragers (Fig. 5A). However, during the conditioning with sugar phase, all three groups showed an increased response to the odour along the trials (Table S1). Once more, pollen foragers from the paired procedure group performed better than both those that underwent the unpaired procedure [PF_PP–PF_UP: Pr(>|z|)=0.018] and nectar foragers [PF_PP–PN_PP: Pr(>|z|)<0.001]. No differences were found between the nectar foragers and the pollen foragers undergoing the unpaired procedure [PF_UP–PN_PP: Pr(>|z|)=0.3087; Fig. 5B].

Fig. 5.

Experiment 1: classical conditioning after pre-conditioning through stimulation of the tarsi with hand-collected kiwi pollen. (A) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the four trials in which the conditioned odour was paired [four reinforced trials, circles for pollen foragers (PF_PP), squares for nectar foragers (NF_PP)] or unpaired [four non-reinforced trials, only for pollen foragers (PF_UP), triangles] with the pollen stimulation (pre-conditioning). (B) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the three trials in which the conditioned odour was paired with a sucrose reward for all groups (conditioning with sugar). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (***P<0.001; *P<0.05).

Fig. 5.

Experiment 1: classical conditioning after pre-conditioning through stimulation of the tarsi with hand-collected kiwi pollen. (A) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the four trials in which the conditioned odour was paired [four reinforced trials, circles for pollen foragers (PF_PP), squares for nectar foragers (NF_PP)] or unpaired [four non-reinforced trials, only for pollen foragers (PF_UP), triangles] with the pollen stimulation (pre-conditioning). (B) Percentage of bees that extended the proboscis as response to the odorant (%PER) during the three trials in which the conditioned odour was paired with a sucrose reward for all groups (conditioning with sugar). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (***P<0.001; *P<0.05).

These results show that although there is a non-measurable unconditioned response during the pre-conditioning, odour–pollen association affected the conditioned response that followed. In most cases, pollen foragers that underwent the paired procedure during pre-conditioning performed better than the other two groups, indicating that both the contingency between pollen and odour as well as forager type are factors that matter.

Experiment 2. Interaction between pollen and sugar reward on PER conditioning

During the paired pollen procedure, bees from both foraging groups showed a high proportion of PER; however, pollen foragers presented better acquisition rates than nectar foragers [PF–NF: Pr(>|z|)=0.046; Fig. 6Ai]. Similarly, this difference extended during the extinction trial, where the extinction rate was lower in pollen foragers than in nectar foragers [PF–NF: Pr(>|z|)=0.017; Fig. 6Ai]. In contrast, during the unpaired pollen procedure, foragers of both groups did not show differences in acquisition [PF–NF: Pr(>|z|)=0.7544; Fig. 6Bi] or extinction [PF–NF: Pr(>|z|)=0.3841; Fig. 6Bii], a result that was found again during the procedure without pollen, where no differences for either acquisition [PF–NF: Pr(>|z|)=0.4622; Fig. 6Ci] or extinction [PF–NF: Pr(>|z|)=0.3396; Fig. 6Cii] were observed.

Fig. 6.

Experiment 2: interaction between pollen and sugar reward on PER conditioning. (A) Paired pollen procedure: percentage of bees that extended the proboscis as response to the odorant (%PER). (B) Unpaired pollen procedure: percentage of bees that extended the proboscis as response to the odorant (%PER). (C) Without pollen procedure: percentage of bees that extended the proboscis as response to the odorant (%PER). Data are shown for five acquisition trials (left) and four extinction trials (right). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (*P<0.05).

Fig. 6.

Experiment 2: interaction between pollen and sugar reward on PER conditioning. (A) Paired pollen procedure: percentage of bees that extended the proboscis as response to the odorant (%PER). (B) Unpaired pollen procedure: percentage of bees that extended the proboscis as response to the odorant (%PER). (C) Without pollen procedure: percentage of bees that extended the proboscis as response to the odorant (%PER). Data are shown for five acquisition trials (left) and four extinction trials (right). Sample sizes are indicated in brackets. Asterisks indicate significant differences in learning performance (*P<0.05).

Experiment 3. Classic olfactory conditioning without reward ingestion

In this experiment, we studied the capacity of bees to establish associations independently of reward ingestion. The results indicate that both nectar and pollen foragers were able to learn along the four conditioning trials (Fig. 7A). Yet, pollen foragers performed better during acquisition than nectar foragers. The analysis detected a significant interaction between trial and group, and revealed that PER levels were different from trial 2 onwards [NF: trial 2=Pr(>|z|)=0.038; trial 3=Pr(>|z|)=0.003; trial 4=Pr(>|z|)=0.016]. Memory retention tests confirmed that non-ingestive reward led to short-term memories in both groups of foragers (Fig. 7B), that did not show significant differences [Pr(>Chi)=0.259]. These results confirm that ingestion is not a critical parameter for reinforcement (Sandoz et al., 2002); however, not contacting their proboscis or letting them ingest the sugar solution drastically reduced the number of bees that achieved the conditioned response.

Fig. 7.

Experiment 3: classical olfactory conditioning without reward ingestion. Percentage of proboscis extension response (%PER) observed during (A) conditioning and (B) a test phase. GLMM: **P<0.01; *P<0.05.

Fig. 7.

Experiment 3: classical olfactory conditioning without reward ingestion. Percentage of proboscis extension response (%PER) observed during (A) conditioning and (B) a test phase. GLMM: **P<0.01; *P<0.05.

In this study, we addressed the rewarding properties of pollen on olfactory associative learning in nectar and pollen honey bee foragers. Interestingly, our results show that the role of pollen as an excitatory stimulus is not applicable to all foragers but is restricted to pollen foragers, as learning performances of both groups were clearly different. Even when in most cases pollen does not elicit a stable unconditioned response (i.e. PER), odour–pollen associations occurred irrespective of behaviour and led to memories that could be traced in conditionings with sugar as the US. Furthermore, we found that pollen foragers learn better than nectar foragers when reinforced without sugar ingestion (pre-ingestive). So far, we detected important differences in how pollen and nectar foragers respond to an odour paired (or not) with pollen, which may contribute to foraging division of labour.

Pollen and the conditioning of the PER

Hand-collected kiwi pollen failed to trigger the extension of the proboscis and, consequently, it failed to olfactory condition the PER. However, it led to a memory that could be traced afterwards in a conditioning with sugar as a reward. Although kiwi flowers are intensely visited by honey bees for their pollen (they do not produce nectar; Jay and Jay, 1984), levels of unconditioned response triggered by kiwi hand-collected pollen contrast with those reported for almond, dandelion and drunken stick in previous studies (Scheiner et al., 2004; Arenas and Farina, 2012), likely owing to differences in the chemical and/or physical composition. In contrast, we found that bee-collected pollen enables the PER conditioning, possibly owing to the small amounts of sugars added by foragers, which might be enough to elicit the extension of the proboscis in bees very sensitive to gustatory stimuli. Levels of the conditioned response were lower than those reported elsewhere (20% versus 70% in Grüter et al., 2008), probably because our procedure only included touching the antennae with intact pollen pellets, whilst the other let the bees ingest, or at least contact, a hydrated pollen paste with their proboscis.

Pollen as reward for pollen but not for nectar foragers

Comparing levels of conditioned response achieved in experiments 1 and 2 (approximately 50% versus 90% of pollen foragers conditioned in three rewarded trials), it is evident that learning performance after pre-conditioning was lower than we would have expected if pollen presented the same releasing function as sugar (e.g. the property to elicit PER). In general, we observed that pollen pre-conditioning impaired acquisition during conditioning. Then, resistance to acquisition was consistent with inhibitory conditioning (Rescorla, 1971; Hellstern and Hammer, 1994; Hellstern et al., 1998), a phenomenon observed if individuals are trained in paired trials of CS and US after the exposure to unpaired presentations of CS and US. As the strength of the link between the odour and the absence of reward in a first training phase determines the resistance to acquisition in a second phase with pairings of CS and US, inhibitory conditioning is expected to be substantial in foragers undergoing the unpaired procedure (i.e. intermediate to long US–CS intervals) but subtle or absent in foragers undergoing the paired procedure. Our findings partially matched this prediction, as pollen foragers were inhibited to a much lesser extent after paired than unpaired pre-conditioning, suggesting that the odour and pollen were associated according to their predictive relationship (Wagner, 1981; Wagner and Larew, 1985) and that odour–pollen pairing prevented, at least in part, the inhibition of the conditioned response by means of excitatory processes triggered by the pollen presentation. Finally, results also show that pollen stimulation alone could not sensitize responses to the odour (i.e. change response as a consequence of the presentation of just one stimulus; Hammer et al., 1994), as unpaired presentations of the odour and pollen could not lead to the same levels of conditioned response we observed in pollen foragers under the paired procedures.

Inhibition of the conditioned response to the CS is also expected if the stimulus we assumed was acting as reward has no reinforcing properties or has a very low reinforcing value. This could be especially the case for nectar foragers, which, even pre-conditioned under a paired procedure, showed a strong resistance to acquisition, suggesting that they did not perceive pollen as an excitatory stimulus. Similar results were observed by Nicholls and Hempel de Ibarra (2013), after they pre-conditioned non-pollen foragers to an odour with a 30% pollen (w/w) solution.

If nectar foragers are ‘insensitive’ or at least not very sensitive to pollen stimulation, the pre-conditioning would turn into an olfactory pre-exposure that leads to latent inhibition, a form of non-associative learning in which the pre-exposure of a single non-reinforced stimulus leads to a strong reduction in learning (Abramson and Bitterman, 1986; Chandra et al., 2000). Differences in latent inhibition have been linked to individual variation in foraging behaviour (Cook et al., 2019), and have been observed to be stronger in pollen than in nectar foragers (Latshaw and Smith, 2005), indicating that improved performances reported here are less likely to be caused by differences in the predisposition to ignore odours.

The hypothesis that nectar foragers are less sensitive than pollen foragers to excitatory cues of pollen is supported by the results of experiment 2, where we used the dual rewarding procedure. There, we observed that pollen foragers performed better than nectar foragers in both the acquisition and extinction phases of the conditioning, if they received the simultaneous presentation of hand-collected pollen and sucrose solution as reinforcement, but not when conditioned with sucrose alone. Although pollen foragers are expected to learn faster and retain memories better than nectar foragers (Scheiner et al., 2001a,b), we managed (by capturing bees in a feeder that offered a solution of low sucrose concentration; see Materials and Methods) to obtain nectar foragers whose performances matched those of pollen foragers when using a 30% sucrose solution as reward (Fig. 6C). Differences between foraging groups are in line with increased learning performance in pollen foragers, presumably owing to an additive effect of the dual sensory input. However, we cannot rule out the possibility that pollen itself, or the way it was delivered to the tarsi of the bees, disturbed nectar foragers in a way that slightly impaired the association. Further experiments including controls for tactile stimulation, but also using intermediate or low sugar concentrations as a reward, will enable us to better understand the pollen–sucrose interaction as reward.

Do pollen and nectar reinforce the same processing pathways?

Olfactory memories established with pollen cannot be directly retrieved on the PER paradigm (present study; Arenas and Farina, 2012). However, pollen and nectar might still be processed within the same pathways. Like in fully satiated bees, which do not extend the proboscis to sugar yet associate an odour with the sucrose US (Hammer and Menzel, 1995), pairing of odour and pollen might occur irrespective of behavioural responses to the odour and the reward. Then, odours learnt with pollen might share the same reinforcement processing pathways as those involved in sugar processing, but they might not include motor and premotor neurons that generate the protrusion of the proboscis (Rehder, 1989). For example, VUMmx1 (ventral unpaired median neuron of the maxillary neuromere 1) responds with long-lasting spike activity to sugar stimulations in the antennae and the proboscis, yet it is not directly involved in generating the PER (Hammer and Menzel, 1995).

Ventral unpaired median neurons such as VUMmx1 might also mediate reinforcement with pollen, as they connect neuropiles such as the antennal lobes, the calyces and the lateral horns (all involved in the olfactory processing pathway) with the suboesophageal ganglion (Hammer and Menzel, 1995; Perry and Barron, 2013). Because the suboesophageal ganglion is located in the ventral nerve cord, between the brain and the thoracic and abdominal ganglia, it might serve as a relay centre for information descending and ascending along the ventral nerve cord, which could be important for both the assessment of pollen with their tarsi as well as the control and coordination of leg movements during pollen gathering. It would be intriguing to test whether VUMmx1, or any other VUM neurons, responds to pollen stimulus to uncover whether pollen and nectar share the same reward-processing pathways.

In contrast, if pollen does not affect the same reward pathway that processes sugar reinforcement, it could still modulate the extension of the proboscis, not directly, but via the olfactory pathway (Hammer and Menzel, 1995; Menzel, 1999). We hypothesize that pollen–odour association alters odour representation in the bee brain, for example, making it more salient or easy to detect or discriminate, which might positively affect olfactory PER conditioning via odour perception.

Different perception and assessment of pollen between pollen and nectar foragers

If we consider that pollen could be perceived as a reward for pollen foragers, it is imperative to determine what bees learn from pollen. Unfortunately, evidence that honey bees can detect compounds present in pollen such as proteins, amino acids, lipids or fatty acids is limited (Kim and Smith, 2000; Pernal and Currie, 2001, 2002; Arenas and Farina, 2012), yet it is likely that nutritional (Pernal and Currie, 2002) and/or non-nutritional pollen components (Lepage and Boch, 1968; Schmidt, 1985) could be perceived as reinforcement during learning. Studies that used pollen-related cues as CS on PER conditioning (Cook et al., 2005; Ruedenauer et al., 2018; Pietrantuono et al., 2019) determined that honey bees can perceive both chemo-tactile and olfactory cues of pollen, and might use them to differentiate between pollen types. So far, very little is known about detection of pollen constituents that excite pollen foragers. Only 10 gustatory receptor genes have been identified in A.mellifera, which is relatively low compared with the 23 gustatory receptor genes detected in the bumblebee Bombus terrestris (Sadd et al., 2015), which exhibited proven abilities to select pollen of higher quality through individual assessment of chemo-tactile cues (Ruedenauer et al., 2015, 2019).

Our results suggest that in tarsi of both front and hind legs, pollen foragers have receptors that detect pollen-related cues and nutrients. Comparing pollen foragers stimulated in different appendages (experiment 1), we found that the delivery of pollen on the tarsi of hind legs performed slightly better during conditioning than when delivered on the antennae (Fig. S1), which contrasts to what has been observed regarding sugar sensitivity: tarsi sensitivity is lower than in antennae (Takeda, 1961; Bitterman et al., 1983; de Brito Sanchez et al., 2008; de Brito Sanchez, 2011; Marshall, 1934). It would be very interesting to search for candidate receptors, in both antennae and tarsi of pollen and nectar foragers, and correlate their presence or density with the bee's predisposition to collect one of the two resource types (Riveros and Gronenberg, 2010; Russell et al., 2017).

The hypothesis that pollen triggers excitatory processes in pollen foragers, whilst it is perceived as a neutral stimulus by nectar foragers, is in line with previous experiments that correlate gustatory responsiveness (Page et al., 1995) with foraging specialization (Page and Fondrk, 1998; Waddington et al., 1998; Pankiw and Page, 2000; Arenas and Kohlmaier, 2019). Furthermore, our findings (experiment 3) support the hypothesis that pollen foragers are more sensitive to pre-ingestive sucrose stimulation than nectar foragers, which agrees with the fact that bees with low response thresholds are less demanding regarding the reward (Scheiner et al., 2004), so they learn faster and retain memories better than bees with low responsiveness (Scheiner et al., 2001a,b). Impaired learning performance with pollen as reward suggests that nectar foragers, in addition to their predisposition to visiting sources producing highly concentrated sugar solutions, are also inefficient at exploiting pollen sources. The opposite might occur in pollen foragers, which, being able to better evaluate the resources through sole antennal or tarsal stimulation, are also attracted to pollen sources as it reinforced their responses and associated cues. The improved learning performance of pollen foragers might be important for the acquisition of relevant environmental cues (e.g. odours) predicting the presence of pollen, thus increasing foraging efficiency. Overall, differences in the perception and learning of pollen within the foraging caste might strongly contribute to colony-level regulation of foraging.

Whether foraging specialization between nectar and pollen foragers in other social bees is driven in the same way as in honey bees is currently less well understood (Russell et al., 2017). Whilst in the stingless bee Tetragonisca angustula it has been recently found that foragers' threshold for sucrose was higher in non-pollen than in pollen foragers (Balbuena and Farina, 2020), no differences were detected between bumblebees (Bombus terrestris) that specialized in pollen and nectar foraging, although the former did learn better in the olfactory PER task (Smith, 2016). Although the mechanisms underlying task specialization may not necessarily be the same for all groups of social bees, we show once more that perception and cognition are relevant for the coordination of the collective foraging response.

We thank W. Farina for the fruitful comments at the early stage of this paper, and M. J. Corriale for help with statistical analyses. We also thank the two anonymous reviewers for their positive comments and suggestions.

Author contributions

Conceptualization: A.A.; Methodology: D.N., E.M., A.A.; Formal analysis: D.N., E.M.; Investigation: D.N., E.M., A.A.; Resources: A.A.; Writing - original draft: D.N., E.M., A.A.; Writing - review & editing: D.N., E.M., A.A.; Supervision: A.A.; Project administration: A.A.; Funding acquisition: A.A.

Funding

This study was partly supported by Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) and by grants from Agencia Nacional de Promoción Científica y Tecnológica (PICT_2017-2688) and University of Buenos Aires CONICET to A.A.

Abramson
,
C. I.
and
Bitterman
,
M. E.
(
1986
).
Latent inhibition in honeybees
.
Anim. Learn. Behav.
14
,
184
-
189
.
Arenas
,
A.
and
Farina
,
W. M.
(
2012
).
Learned olfactory cues affect pollen-foraging preferences in honeybees, Apis mellifera
.
Anim. Behav.
83
,
1023
-
1033
.
Arenas
,
A.
and
Farina
,
W. M.
(
2014
).
Bias to pollen odors is affected by early exposure and foraging experience
.
J. Insect Physiol.
66
,
28
-
36
.
Arenas
,
A.
and
Kohlmaier
,
M. G.
(
2019
).
Nectar source profitability influences individual foraging preferences for pollen and pollen-foraging activity of honeybee colonies
.
Behav. Ecol. Sociobiol.
73
,
34
.
Balbuena
,
M. S.
and
Farina
,
W. M.
(
2020
).
Chemosensory reception in the stingless bee Tetragonisca angustula
.
J. Insect Physiol.
125
,
104076
.
Bates
,
D.
,
Mächler
,
M.
,
Bolker
,
B. M.
and
Walker
,
S. C.
(
2015
).
Fitting linear mixed-effects models using lme4
.
J. Stat. Softw.
67
,
1-48
.
Beshers
,
S. N.
and
Fewell
,
J. H.
(
2001
).
Models of division of labor in social insects
.
Annu. Rev. Entomol.
46
,
413
-
440
.
Bitterman
,
M.
,
Menzel
,
R.
,
Fietz
,
A.
and
Schäfer
,
S.
(
1983
).
Classical conditioning of proboscis extension in honeybees (Apis mellifera)
.
J. Comp. Physiol.
97
,
107
-
119
.
Bonabeau
,
E.
and
Theraulaz
,
G.
(
1999
).
Role and variability of response thresholds in the regulation of division of labor in insect societies
. In
Information Processing in Social Insects
, (ed.
C.
Detrain
,
J. L.
Deneubourg
and
J. M.
Pasteels
) pp.
141
-
163
.
Basel
:
Birkhäuser
.
Chandra
,
S. B. C.
,
Hosler
,
J. S.
and
Smith
,
B. H.
(
2000
).
Heritable variation for latent inhibition and its correlation with reversal learning in honeybees (Apis mellifera)
.
J. Comp. Psychol.
114
,
86
-
97
.
Cook
,
S. M.
,
Sandoz
,
J.-C.
,
Martin
,
A. P.
,
Murray
,
D. A.
,
Poppy
,
G. M.
and
Williams
,
I. H.
(
2005
).
Could pollen odours play a role in the foraging behaviour of the honey bee (Apis mellifera)
.
Physiol. Entomol.
30
,
164
-
174
.
Cook
,
C. N.
,
Mosqueiro
,
T.
,
Brent
,
C. S.
,
Ozturk
,
C.
,
Gadau
,
J.
,
Pinter-Wollman
,
N.
and
Smith
,
B. H.
(
2019
).
Individual differences in learning and biogenic amine levels influence the behavioural division between foraging honey bee scouts and recruits
.
J. Anim. Ecol.
88
,
236
-
246
.
Couzin
,
D.
(
2009
).
Collective cognition in animal groups
.
Trends. Cognit. Sci.
13
,
36
-
43
.
Couzin
,
I. D.
,
Ioannou
,
C. C.
,
Demirel
,
G.
,
Gross
,
T.
,
Torney
,
C. J.
,
Hartnett
,
A.
,
Conradt
,
L.
,
Levin
,
S. A.
and
Leonard
,
N. E.
(
2011
).
Uninformed individuals promote democratic consensus in animal groups
.
Science
334
,
1578
-
1580
.
de Brito Sanchez
,
M. G.
(
2011
).
Taste perception in honey bees
.
Chem. Senses
36
,
675
-
692
.
de Brito Sanchez
,
M. G.
,
Chen
,
C.
,
Li
,
J.
,
Liu
,
F.
,
Gauthier
,
M.
and
Giurfa
,
M.
(
2008
).
Behavioral studies on tarsal gustation in honeybees: sucrose responsiveness and sucrose-mediated olfactory conditioning
.
J. Comp. Physiol. A
194
,
861
-
869
.
Duarte
,
A.
,
Weissing
,
F. J.
,
Pen
,
I.
and
Keller
,
L.
(
2011
).
An evolutionary perspective on self-organized division of labor in social insects
.
Annu. Rev. Ecol. Evol. Syst.
42
,
91
-
110
.
Dukas
,
R.
(
1998
).
Cognitive Ecology: The Evolutionary Ecology of Information Processing and Decision Making
.
University of Chicago Press
.
Grüter
,
C.
,
Arenas
,
A.
and
Farina
,
W. M.
(
2008
).
Does pollen function as a reward for honeybees in associative learning?
Insects Soc.
55
,
425
-
427
.
Hammel
,
B.
,
Vollet-Neto
,
A.
,
Menezes
,
C.
,
Nascimento
,
F. S.
,
Engels
,
W.
and
Grüter
,
C.
(
2016
).
Soldiers in a stingless bee: work rate and task repertoire suggest they are an elite force
.
Am. Nat.
187
,
120
-
129
.
Hammer
,
M.
and
Menzel
,
R.
(
1995
).
Learning and memory in the honeybee
.
J. Neurosci.
15
,
1617
-
1630
.
Hammer
,
M.
,
Braun
,
G.
and
Mauelshagen
,
J.
(
1994
).
Food-induced arousal and nonassociative learning in honeybees: dependence of sensitization on the application site and duration of food stimulation
.
Behav. Neural Biol.
62
,
210
-
223
.
Hellstern
,
F.
and
Hammer
,
M.
(
1994
).
Backward inhibitory learning in honeybees
. In
Sensory transduction. Proceedings of the 22d Göttingen Neurobiology Conference
(ed.
N.
Elsner
and
H.
Breer
), pp.
827
.
Stuttgart
:
Thieme
.
Hellstern
,
F.
,
Malaka
,
R.
and
Hammer
,
M.
(
1998
).
Backward inhibitory learning in honeybees: a behavioral analysis of reinforcement processing
.
Learn. Mem.
4
,
429
-
444
.
Hölldobler
,
B.
and
Wilson
,
E.
(
2008
).
The Superorganism: The Beauty, Elegance, and Strangeness of Insect Societies
.
W.W. Norton & Company
.
Jay
,
D.
and
Jay
,
C.
(
1984
).
Observations of honeybees on Chinese gooseberries (‘kiwifruit’) in New Zealand
.
Bee World
65
,
155
-
166
.
Kim
,
Y. S.
and
Smith
,
B. H.
(
2000
).
Effect of an amino acid on feeding preferences and learning behavior in the honey bee, Apis mellifera
.
J. Insect Physiol.
46
,
793
-
801
.
Kuwabara
,
M.
(
1957
).
Bildung Des Bedingten reflexes Von Pavlovs Typus Bei Der Honigbiene, Apis mellifera
.
J. Fac. Sci. Hokkaido Univ. Zool.
13
,
458
-
464
.
Latshaw
,
J. S.
and
Smith
,
B. H.
(
2005
).
Heritable variation in learning performance affects foraging preferences in the honey bee (Apis mellifera)
.
Behav. Ecol. Sociobiol.
58
,
200
-
207
.
Lenth
,
R. V.
(
2015
).
The case against normal plots of effects
.
J. Qual. Technol.
47
,
91
-
97
.
Lepage
,
M.
and
Boch
,
R.
(
1968
).
Pollen lipids attractive to honeybees
.
Lipids
3
,
530
-
534
.
Marshall
,
J.
(
1934
).
On the sensitivity of the chemoreceptors on the antenna and fore-tarsus of the honey-bee
.
J. Exp. Biol.
12
,
17
-
26
.
Mattiacci
,
A.
(
2019
).
Conducta de recolección de alimento en Vespula germanica: implicancias para el control de avispas sociales invasoras en la Patagonia
.
PhD thesis
,
Universidad de Buenos Aires,Facultad de Ciencias Exactas y Naturales
.
Menzel
,
R.
(
1999
).
Memory dynamics in the honeybee
.
J. Comp. Physiol. A Sensory, Neural, Behav. Physiol.
185
,
323
-
340
.
Nicholls
,
E.
and
Hempel de Ibarra
,
N.
(
2013
).
Pollen elicits proboscis extension but does not reinforce PER learning in honeybees
.
Insects
4
,
542
-
557
.
Oster
,
G. F.
and
Wilson
,
E. O.
(
1978
).
Caste and Ecology in the Social Insects
.
Princeton University Press
.
Page
,
R. E.
, Jr
and
Fondrk
,
M. K.
(
1998
).
The effect of genotype on response thresholds to sucrose and foraging behavior of honey bees (Apis mellifera L.)
.
J. Comp. Physiol. A.
182
,
489
-
500
.
Page
,
R.
,
Robinson
,
G.
,
Fondrk
,
M.
and
Nasr
,
M.
(
1995
).
Effects of worker genotypic diversity on honey bee colony development and behavior (Apis mellifera L.)
.
Behav. Ecol. Sociobiol.
36
,
387
-
396
.
Page
,
R. E.
,
Scheiner
,
R.
,
Erber
,
J.
and
Amdam
,
G. V.
(
2006
).
The development and evolution of division of labor and foraging specialization in a social insect (Apis mellifera L.)
.
Curr. Top. Dev. Biol.
74
,
253
-
286
.
Pankiw
,
T.
and
Page
,
R. E.
(
2000
).
Response thresholds to sucrose predict foraging division of labor in honeybees
.
Behav. Ecol. Sociobiol.
47
,
265
-
267
.
Perez
,
M.
,
Rolland
,
U.
,
Giurfa
,
M.
and
d'Ettorre
,
P.
(
2013
).
Sucrose responsiveness, learning success, and task specialization in ants
.
Learn. Mem.
20
,
417
-
420
.
Pernal
,
S. F.
and
Currie
,
R. W.
(
2001
).
The influence of pollen quality on foraging behavior in honeybees (Apis mellifera L.)
.
Behav. Ecol. Sociobiol.
51
,
53
-
68
.
Pernal
,
S. F.
and
Currie
,
R. W.
(
2002
).
Discrimination and preferences for pollen-based cues by foraging honeybees, Apis mellifera L
.
Anim. Behav.
63
,
369
-
390
.
Perry
,
C. J.
and
Barron
,
A. B.
(
2013
).
Honey bees selectively avoid difficult choices
.
Proc. Natl. Acad. Sci. USA
110
,
19155
-
19159
.
Pietrantuono
,
A. L.
,
Requier
,
F.
,
Fernández-Arhex
,
V.
,
Winter
,
J.
,
Huerta
,
G.
and
Guerrieri
,
F.
(
2019
).
Honeybees generalize among pollen scents from plants flowering in the same seasonal period
.
J. Exp. Biol.
222
,
jeb201335
.
Raguso
,
R. A.
(
2008
).
Wake up and smell the roses: the ecology and evolution of floral scent
.
Annu. Rev. Ecol. Evol. Syst.
39
,
549
-
569
.
Rehder
,
V.
(
1989
).
Sensory pathways and motoneurons of the proboscis reflex in the suboesophageal ganglion of the honey bee
.
J. Comp. Neurol.
279
,
499
-
513
.
Rescorla
,
R. A.
(
1971
).
Summation and retardation tests of latent inhibition
.
J. Comp. Physiol. Psychol.
75
,
77
-
81
.
Riveros
,
A. J.
and
Gronenberg
,
W.
(
2010
).
Sensory allometry, foraging task specialization and resource exploitation in honeybees
.
Behav. Ecol. Sociobiol.
64
,
955
-
966
.
Robinson
,
G. E.
(
1992
).
Regulation of division of labor in insect societies
.
Annu. Rev. Entomol.
37
,
637
-
665
.
Robinson
,
G. E.
and
Page
,
R. E.
(
1989
).
Genetic basis for division of labor in an insect society
. In
The Genetics of Social Evolution
(ed.
M. D.
Breed
and
R. E.
Page
), pp.
61
-
80
.
Boulder, CO
:
Westview Press Inc
.
Ruedenauer
,
F. A.
,
Spaethe
,
J.
and
Leonhardt
,
S. D.
(
2015
).
How to know which food is good for you: bumblebees use taste to discriminate between different concentrations of food differing in nutrient content
.
J. Exp. Biol
218
,
2233
-
2240
.
Ruedenauer
,
F. A.
,
Wöhrle
,
C.
,
Spaethe
,
J.
and
Leonhardt
,
S. D.
(
2018
).
Do honeybees (Apis mellifera) differentiate between different pollen types?
PLoS ONE
13
,
e0205821
.
Ruedenauer
,
F. A.
,
Leonhardt
,
S. D.
,
Lunau
,
K.
and
Spaethe
,
J.
(
2019
).
Bumblebees are able to perceive amino acids via chemotactile antennal stimulation
.
J. Comp. Physiol. A
205
,
321
-
331
.
Russell
,
A. L.
,
Morrison
,
S. J.
,
Moschonas
,
E. H.
and
Papaj
,
D. R.
(
2017
).
Patterns of pollen and nectar foraging specialization by bumblebees over multiple timescales using RFID
.
Sci. Rep.
7
,
1
-
13
.
Sadd
,
B. M.
,
Barribeau
,
S. M.
,
Bloch
,
G.
,
De Graaf
,
D. C.
,
Dearden
,
P.
, Elsik,C. G.,
Gadau
,
J.
,
Grimmelikhuijzen
,
C. J.
,
Hasselmann
,
M.
and
Lozier
,
J. D.
(
2015
).
The genomes of two key bumblebee species with primitive eusocialorganization
.
Genome Biol. 16, 1-32
.
Sandoz
,
J. C.
,
Hammer
,
M.
and
Menzel
,
R.
(
2002
).
Side-specificity of olfactory learning in the honeybee: US input side
.
Learn. Mem.
9
,
337
-
348
.
Scheiner
,
R.
,
Page
,
R. E.
and
Erber
,
J.
(
2001a
).
The effects of genotype, foraging role, and sucrose responsiveness on the tactile learning performance of honey bees (Apis mellifera L.)
.
Neurobiol. Learn. Mem.
76
,
138
-
150
.
Scheiner
,
R.
,
Page
,
R. E.
and
Erber
,
J.
(
2001b
).
Responsiveness to sucrose affects tactile and olfactory learning in preforaging honey bees of two genetic strains
.
Behav. Brain Res.
120
,
67
-
73
.
Scheiner
,
R.
,
Page
,
R. E.
and
Erber
,
J.
(
2004
).
Sucrose responsiveness and behavioral plasticity in honey bees (Apis mellifera)
.
Apidologie
35
,
133
-
142
.
Scheiner
,
R.
,
Kuritz-Kaiser
,
A.
,
Menzel
,
R.
and
Erber
,
J.
(
2005
).
Sensory responsiveness and the effects of equal subjective rewards on tactile learning and memory of honeybees
.
Learn. Mem.
12
,
626
-
635
.
Schmidt
,
J. O.
(
1985
).
Phagostimulants in pollen
.
J. Apic. Res.
24
,
107
-
114
.
Seeley
,
T. D.
(
2009
).
The Wisdom of the Hive: the Social Physiology of Honey Bee Colonies
.
Harvard University Press
.
Sih
,
A.
and
Del Giudice
,
M.
(
2012
).
Linking behavioural syndromes and cognition: a behavioural ecology perspective
.
Philos. Trans. R. Soc. Lond. B Biol. Sci.
367
,
2762
-
2772
.
Smith
,
K. E.
(
2016
).
Causes and consequences of variation in learning performance in the bumblebee (Bombus terrestris)
.
PhD thesis
,
University of London
,
Royal Holloway
.
Takeda
,
K.
(
1961
).
Classical conditioned response in the honey bee
.
J. Insect Physiol.
6
,
168
-
179
.
Waddington
,
K.
,
Nelson
,
C.
and
Page
,
R.
(
1998
).
Effects of pollen quality and genotype on the dance of foraging honey bees
.
Anim. Behav.
56
,
35
-
39
.
Wagner
,
A. R.
(
1981
).
SOP: a model of automatic memory processing in animal behavior
. In
Information Processing in Animals: Memory Mechanisms
(ed.
N.
Spear
and
R.
Miller
), pp.
5
-
48
.
Hillsdale, NJ
:
Lawrence Erlbaum Associates, Inc
.
Wagner
,
A. R.
and
Larew
,
M. B.
(
1985
).
Opponent processes and Pavlovian inhibition
. In
Information Processing in Animals: Conditioned Inhibition
(ed.
R. R.
Miller
and
N. E.
Spear
), pp.
223
-
266
.
Hillsdale, NJ
:
Lawrence Erlbaum Associates, Inc
.
Wilson
,
E. O.
(
1971
).
The Insect Societies
.
Harvard University Press
.
Wilson
,
E. O.
(
1987
).
Causes of ecological success: the case of the ants
.
J. Anim. Ecol.
56
,
1
-
9
.
Winston
,
M. L.
(
1987
).
The Biology of the Honey Bee
.
Cambridge, MA
:
Harvard University Press
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information