Olfactory subsystems in the honeybee: sensory supply and sex specificity
The antennae of honeybee (Apis mellifera) workers and drones differ in various aspects. One striking difference is the presence of Sensilla basiconica in (female) workers and their absence in (male) drones. We investigate the axonal projection patterns of olfactory receptor neurons (ORNs) housed in S. basiconica in honeybee workers by using selective anterograde labeling with fluorescent tracers and confocal-microscopy analysis of axonal projections in antennal lobe glomeruli. Axons of S. basiconica-associated ORNs preferentially projected into a specific glomerular cluster in the antennal lobe, namely the sensory input-tract three (T3) cluster. T3-associated glomeruli had previously been shown to be innervated by uniglomerular projection (output) neurons of the medial antennal lobe tract (mALT). As the number of T3 glomeruli is reduced in drones, we wished to determine whether this was associated with the reduction of glomeruli innervated by medial-tract projection neurons. We retrogradely traced mALT projection neurons in drones and counted the innervated glomeruli. The number of mALT-associated glomeruli was strongly reduced in drones compared with workers. The preferential projections of S. basiconica-associated ORNs in T3 glomeruli together with the reduction of mALT-associated glomeruli support the presence of a female (worker)-specific olfactory subsystem that is partly innervated by ORNs from S. basiconica and is associated with the T3 cluster of glomeruli and mALT projection neurons. We propose that this olfactory subsystem supports parallel olfactory processing related to worker-specific olfactory tasks such as the coding of colony odors, colony pheromones and/or odorants associated with foraging on floral resources.
KeywordsSensilla basiconica Antennal lobe Glomeruli Projection neuron Honeybee drone
Olfaction is an ancient sensory modality and plays a crucial role in most animals for approaching or avoiding various odor sources and for judging their quality in a variety of behavioral contexts. Whereas odorant reception at the molecular level exhibits distinct differences between vertebrates and insects, the basic wiring pattern of receptor neurons with second-order neurons within the primary olfactory centers, the vertebrate olfactory bulb and the insect antennal lobe (AL) shows several striking similarities. These have been the subject of intense research over recent years (for reviews, see Hildebrand and Shepherd 1997; Ache and Young 2005; Touhara and Vosshall 2009; Martin et al. 2011).
Insect antennae are covered with various types of sensory sensilla; most of them being specialized for chemoreception but also for hygro-, mechano- and thermoreception (Esslen and Kaissling 1976; Maronde 1991; Ai et al. 2007). Olfactory sensilla house the olfactory receptor neurons (ORNs) that extend axons into spheroidal structures termed glomeruli to form synaptic connections with local interneurons and projection neurons (PNs). Glomeruli represent the functional units of the AL (e.g., Anton and Homberg 1999). In the honeybee, Sensilla placodea, Sensilla trichoidea and Sensilla basiconica have been classified as olfactory sensilla, either according to their odor-response profiles in single-sensillum recordings (S. placodea) or based on specific anatomical features (S. trichoidea, S. basiconica; Lacher 1964; Esslen and Kaissling 1976; Akers and Getz 1993; Getz and Akers 1993). ORN axons from olfactory sensilla project via four distinct AL sensory-input tracts to four clusters of glomeruli in the AL termed the T1-T4 cluster (Mobbs 1982; Galizia et al. 1999; Abel et al. 2001; Kirschner et al. 2006). Axons from ORNs in S. placodea project to all four clusters of glomeruli T1-T4 (Brockmann and Brückner 1995; Kelber et al. 2006) and single-sensillum recordings from S. placodea have revealed responses to a broad range of odorants (Getz and Akers 1993). This might be either caused by the broad tuning of ORNs or attributable to the finding that S. placodea house many individual ORNs (between 7 and 23; Kelber et al. 2006), each covering a certain spectrum of molecular receptive ranges.
In insects, PNs convey the olfactory information to the mushroom bodies (MBs), higher sensory association centers and sites associated with learning and memory (Gerber et al. 2004; Davis 2005; Giurfa 2007; Hourcade et al. 2010; Cervantes-Sandoval et al. 2013). Uniglomerular PNs in the honeybee and other Hymenoptera have been shown to project to the MBs and lateral horn (LH) via two parallel tracts: the medial and the lateral AL tracts (mALT and lALT) forming a dual olfactory pathway (Abel et al. 2001; Kirschner et al. 2006; Galizia and Rössler 2010; Rössler and Brill 2013; tract terminology according to Ito et al. 2014). Comparative anatomical studies indicate that a dual olfactory pathway probably emerged in the basal Hymenoptera (Dacks and Nighorn 2011; Rössler and Zube 2011; Rössler and Brill 2013). However, the selective pressure that promoted the evolution of a dual olfactory pathway within this group of insects remains to be further investigated (Rössler and Brill 2013).
Female honeybee workers show complex social behavior that is largely influenced by pheromonal communication (Slessor et al. 2005; Le Conte and Hefetz 2008). This is different in male drones, which mainly perform reproductive tasks, do not forage actively for food and might not need to distinguish minor changes in colony pheromone concentrations. On the other hand, drones are highly sensitive to the queen sex-pheromone (Gary 1962). Therefore, differences in the olfactory system reflecting these behavioral specializations are likely to exist between honeybee workers and drones. One striking sex-specific difference is the absence of S. basiconica on drone antennae (Esslen and Kaissling 1976; Nishino et al. 2009). Furthermore, drone ALs contain a smaller number of glomeruli compared with both female castes (workers and queens) but comprise several enlarged macroglomeruli (Arnold et al. 1985; Sandoz 2006; Groh and Rössler 2008; Nishino et al. 2009). The largest macroglomerulus has been shown to respond to the major component of the queen mandibular pheromone (Sandoz 2006; Wanner et al. 2007). The reduction of AL glomeruli is mostly associated with the T3 cluster (Nishino et al. 2009), which has been demonstrated to be mainly innervated by medial tract PNs in honeybee workers (Kirschner et al. 2006). Comparative studies in other Hymenoptera indicate that the lack of S. basiconica in males is a characteristic trait across both social and solitary Hymenoptera (Ågren 1977, 1978; Wcislo 1995; Ågren and Hallberg 1996; Nakanishi et al. 2009, 2010; Mysore et al. 2010; Nishikawa et al. 2012; Streinzer et al. 2013). In the leaf-cutting ant Atta vollenweideri, S. basiconica have been found exclusively to innervate a specific (T6) cluster of AL glomeruli (Kelber et al. 2010).
The absence of S. basiconica, together with the reduction of glomeruli in the T3 cluster, in honeybee drones suggests that ORNs from S. basiconica preferentially innervate glomeruli in the T3 cluster and are associated with medial-tract PNs. To test this hypothesis, we investigated the axonal projections of ORNs in the hair-like olfactory antennal sensilla of female worker bees, with a special focus on S. basiconica and, in particular, their glomerular innervation patterns and their association with PN output tracts. Furthermore, we retrogradely labeled the mALT in drones to analyze whether glomeruli associated with this tract are reduced.
Materials and methods
Honeybee workers and drones (Apis mellifera carnica) were collected from bee hives of the institutional bee station at the University of Würzburg. Workers were taken from the entrance of the hive and drones were caught directly in the hive. In both cases we did not control for age. The animals were cooled in a refrigerator (4 °C), fixed in custom made plastic holders and provided with sugar solution (40 %) ad libitum.
Staining of axonal projections
Staining of medial tract projection neurons in drones
PNs of the mALT were retrogradely stained by using methods described in detail by Kirschner et al. (2006), Zube et al. (2008) and Rössler and Zube (2011). Head capsules were opened and glands and tracheae covering the brain were removed. The tissue between the vertical lobe of the MB and the AL was punctured with a fine glass electrode on the medial side, where mALT fibers run relatively close to the surface of the brain. A few crystals of either Microruby or Biotin Dextran were gently applied to the punctured tissue by means of a broken micropipette. Mass-staining of all AL output tracts was achieved by using a similar technique and application of dye crystals into the center of the AL. Mass-staining of AL glomeruli was performed by the cutting of the antennae and the application of a droplet of Microruby (about 3-5 % in distilled water) to the antennal nerve stump. In all cases, animals were kept for 4–6 h in the dark under moist conditions to allow the dye to be transported intracellularly.
Single PN staining
For single PN staining, the glial sheath was gently removed from the AL with fine forceps. PN cell bodies were approached with a patch-clamp pipette under visual control by using an upright microscope (BX51WI, Olympus Imaging Europa) and a micromanipulator (Junior Unit, Luigs & Neumann). We employed the intracellular [(K-gluconate (110 mM), HEPES (25 mM), KCl (10 mM), MgCl2 (5 mM), Mg-ATP (3 mM), Na-GTP (0.5 mM), EGTA (0.5 mM), pH 7.2, 284 mOsm)] and extracellular [(NaCl (140 mM), KCl (5 mM), MgCl2 (1 mM), CaCl2 (2.5 mM), NaHCO3 (4 mM), NaH2PO4 (1.2 mM), HEPES (6 mM), glucose (14 mM), pH 7.4, 326 mOsm)] solutions of Palmer et al. (2013). For staining, 0.5-1 % Lucifer Yellow (L0259, Sigma-Aldrich Chemie, Steinheim, Germany) was added to the intracellular solution. Whole cell patch clamp recordings were established by using an Axon system (Axopatch 200B, Digidata 1440A, ClampEx, Molecular Devices, Oregon, USA) and the membrane voltage was kept at -60 mV.
Brains were placed in fixative solution (4 % formaldehyde) overnight, following the staining procedure and then rinsed five times for 10 min in PBS (phosphate-buffered saline, pH 7.2). Biotin-Dextran-injected brains were stained with Alexa-488-conjugated Streptavidin (S-11226, Molecular Probes, Eugene, Ore., USA) in PBS with 0.2% Triton X (1:125) for 48 h. Subsequently, brains were again rinsed five times for 10 min in PBS. Microruby- and Biotin-Dextran-labeled brains were dehydrated in an ascending ethanol series (50, 70, 90, 95, 2× 100 %, each 10 min) and cleared and mounted in methyl salicylate (M2047, Sigma-Aldrich Chemie) for confocal microscopy. Afterwards, brains were scanned with a confocal laser-scanning microscope (Leica TCS SP2; Leica Microsystems, Wetzlar, Germany). Images were processed with AMIRA 3.1.1 and 5.4 software (Mercury Computer Systems, Berlin, Germany) and AL glomeruli in one preparation were reconstructed by using the AMIRA wrapping tool. Image stacks were further processed with ImageJ 1.46j (Wayne Rasband, National Institutes of Health, Befesta, Md., USA). To count the glomeruli in all other successful mALT stainings (that were not three-dimensionally reconstructed in detail), we marked glomeruli in image stacks by using the segmentation editor implemented in ImageJ. Contrast and brightness were adjusted with GIMP 2.8.2 (GNU Image Manipulation Program, http://www.gimp.org) and images in figures were arranged with CorelDrawX6 (Corel, Ottawa, ON, Canada).
Identification of glomeruli
After the successful staining of axonal projections, AL reconstructions in workers were mapped onto a template AL by using the VOI (volume of interest) method described in Kelber et al. (2006). The antennal nerve, the T1 tract and a specific glomerulus (A17; after Galizia et al. 1999) were used as landmarks. Additionally, one AL reconstruction of a mass-staining of ORN axons of the whole antennal nerve was mapped onto the same template AL. This reconstruction allowed the identification of the AL input tracts in the template AL. With the identified glomeruli and the input tracts in the template AL, the matching of sensilla-stained ALs allowed the assignment of axonal projections of ORNs to a specific input tract and to particular glomeruli.
Relative innervation frequencies of the input clusters (T1-T4) were analyzed with a Friedman analysis of variance (ANOVA) and a post-hoc Wilcoxon rank sum test with Bonferroni correction for multiple comparisons. The distribution of ORN axons in the various glomerular clusters originating from the staining of S. basiconica-rich regions was compared with a hypothetical distribution in the glomerular input clusters based on the natural distribution by using Fisher’s exact test. All statistic tests were performed with R 2.10.1 (R Foundation for Statistical Computing, Vienna, Austria).
ORN axon projections from many S. trichoidea and S. basiconica are broadly distributed into AL glomeruli
Axons from ORNs in S. trichoidea of a single antennal segment arborize in the T1, T2 and T3 cluster of glomeruli
Axons from ORNs in S. basiconica preferentially innervate the T3 cluster of glomeruli
Number of glomeruli innervated by mALT projection neurons is reduced in honeybee drones
This study shows that axons of ORNs from S. basiconica, which are absent on drone antennae, mainly project into the T3 cluster of glomeruli in the AL of worker bees (Fig. 7). The T3 cluster of glomeruli has been demonstrated to be reduced in drones (Nishino et al. 2009) and our results reveal that mALT PNs (which, in the female AL, receive input from the T3 cluster) innervate fewer glomeruli in drones compared with workers. This indicates that information from the S. basiconica is associated with the mALT pathway in females and further suggests that this part of the mALT pathway is reduced in drones compared with workers. A similar reduction of the mALT-associated portion of AL glomeruli has been shown in males of the ant Camponotus floridanus (Zube and Rössler 2008).
General features and sorting of axonal projections of ORNs
The trajectories of individual axons in the mass-stainings indicate that ORN axons are guided to their glomerular targets via a two-step sorting process, one at the segregation into T1-T4 at the entrance of the antennal nerve and a second one at the end of the tracts, just before the axons diverge to their glomerular targets. We show that ORN axons from the various antennal segments run in parallel along the antennal nerve until they reach the AL entrance. This is analogous to the ORN axon-sorting zone in the moth Manduca sexta (Rössler et al. 1999). After the initial rearrangements at the AL entrance, however, axonal projections in the honeybee proceed again in parallel along the tracts (T1-T4) until they reach the glomerular clusters. At these glomerular clusters, ORN axons appear to be rearranged in another crossing pattern until they terminate in distinct layers within individual glomeruli. This indicates that the organization in the four olfactory sensory tracts in the honeybee involves a more complex sorting process compared with the single sorting zone at the AL entrance shown in moths (Rössler et al. 1999). This process requires further clarification in future developmental studies.
Within individual glomeruli, ORN axons of sensilla from the distal and proximal segments of the antenna terminate in different layers (Fig. 2). These projection patterns indicate a change from the topographic position of sensilla on the antenna to the spatial organization of ORN target fields within single AL glomeruli. In contrast to the ORN projections, PN dendrites might arborize throughout the entire glomerulus (Fig. 2c, d; see also Müller et al. 2002; Krofczik et al. 2008) indicating that individual PNs receive convergent input from ORNs from widely separated topographical locations that innervate the same glomerulus.
Glomerular targets of ORN axons from S. trichoidea
Selective mass-staining of S. trichoidea labels bundles of ORN axons projecting to the AL (Fig. 3). The trajectories of ORN axons can be followed along three tracts (T1, T2, T3; Fig. 3). In none of the 13 preparations have we found staining of T4 glomeruli (Fig. 3d). Interestingly, despite a wide distribution of axons in almost all glomeruli of the AL, slightly brighter staining has been detected in T3 glomeruli compared with T1 glomeruli. Whether this is attributable to a higher density and/or brighter labeling of individual axons cannot be resolved with our method. ORN axons from S. placodea have also been shown to project into all four glomerular clusters in the honeybee (Kelber et al. 2006). For a more detailed analysis, single fills of S. trichoidea ORNs would be necessary. Our staining demonstrates that ORN projections from S. trichoidea from only one antennal segment are widely distributed across three glomerular clusters (T1-T3) that contain >95 % of all AL glomeruli in the honeybee (Kirschner et al. 2006). Assuming that this is true for all antennal segments (we found similar results for segments 3, 5, 6, 7 and 8), this implies a high redundancy of the antennal segments within individual glomeruli. This ensures that the antenna efficiently captures sufficient odor molecules for odor identification but rather excludes a topographical resolution of odor reception (e.g., ORNs at the proximal or distal part of the antenna). In contrast to ORNs, PNs have dendritic arborizations across entire glomeruli (Fig. 2c, d). Whether the position of ORN projections in a distinct layer of the glomerulus (and its position on the antenna) in the honeybee has any physiological impact on PN responses, however, needs to be investigated in physiological recordings.
Projection patterns of ORN axons from S. basiconica
Mass-staining techniques of hair-like sensilla on entire antennal segments have not allowed us to resolve glomerular cluster preferences of S. basiconica-associated ORN axons, as even the S. basiconica-rich segment 9 contains over 10× more S. trichoidea than S. basiconica (Esslen and Kaissling 1976). To increase the proportion of S. basiconica-associated ORNs, we selectively stained hair-like sensilla in the distal regions of individual antennal segments that show a ring-like arrangement of S. basiconica (Fig. 1). These experiments revealed that ORN axons from these sensilla terminate to a significantly higher proportion in glomeruli of the T3 cluster of AL glomeruli (Figs. 5, 7). In these staining experiments, we have preferentially aimed to cut S. basiconica but because of the small size of these sensilla in the honeybee, staining of S. trichoidea, in most cases, cannot be completely excluded. Therefore, we used a statistical method to analyze cluster preference in these selective stainings. Furthermore, the approved staining (by visual inspection) of a single S. basiconicum has shown that nine of the T3 glomeruli and only one T1 glomerulus are innervated thereby supporting the multiple sensilla staining results. The single S. basiconicum stained in our experiments comprised 10 ORNs. This is in agreement with results from histological investigations showing that S. basiconica in the honeybee contain 8-12 ORN dendrites (Nishino et al. 2009). In other Hymenoptera, ORN numbers in S. basiconica have been demonstrated to range between 30-40 ORNs in wasps (Lacher 1964) and 3-53 ORNs in ants (Kelber et al. 2010). Until now, the notion that S. basiconica are indeed olfactory sensilla is mainly based on morphological criteria. Whereas the morphological evidence is strong, however, clear physiological evidence is still lacking. Recordings from S. basiconica have only been mentioned as unpublished results in Akers and Getz (1993). Our anterograde staining of the axonal projections of ORNs housed in S. basiconica has revealed innervations of relatively high numbers of preferentially T3 glomeruli in the AL. This adds further evidence to the assertion that S. basiconica are olfactory sensilla (Slifer and Sekhon 1961; Lacher 1964). Our results suggest that axons of S. basiconica-associated ORNs preferentially project to glomeruli of the T3 cluster and only partially to glomeruli of the T1 cluster.
Differences between honeybee drones and workers and potential functional consequences
Interestingly, S. basiconica are present in honeybee workers and queens but not in drones (Esslen and Kaissling 1976). The absence of S. basiconica in males seems to be a typical feature in Hymenoptera including the honeybee, bumblebees, solitary bees and ants (Ågren 1977; 1978; Wcislo 1995; Ågren and Hallberg 1996; Nakanishi et al. 2009, 2010; Mysore et al. 2010; Nishikawa et al. 2012; Streinzer et al. 2013). Wcislo (1995) argued that drones have reduced the number of a certain type of olfactory sensilla to increase their capabilities for pheromone detection and thereby to increase their probability for encountering a queen faster than other drones. The finding that drones have extremely high numbers of pore plates (S. placodea), which also house the pheromone-specific sensilla (Kaissling and Renner 1968), supports this view. Three of the enlarged glomeruli (macroglomeruli) are located in the T1 cluster of glomeruli (Nishino et al. 2009) and calcium-imaging studies have shown that components of the queen mandibular pheromone are processed in one of these enlarged glomeruli (Sandoz 2006). The processing of pheromone information in macroglomeruli is well known from sex pheromone communication in various species of moths (e.g., Christensen and Hildebrand 1987; Anton and Hansson 1994; Greiner et al. 2004) and has also been shown for an enlarged glomerulus which processes components of the trail pheromone in large leaf-cutting ant workers (Kleineidam et al. 2005; Kelber et al 2009; Kuebler et al 2010). The reduced number of S. trichoidea and the complete absence of S. basiconica and the associated neuronal pathway in the AL of drones might allow higher capacities for queen-pheromone-specific sensilla and a larger sex-pheromone-processing neuronal circuitry in the AL (Fig. 7).
Female-specific olfactory subsystems and their possible function
We have been able to demonstrate that the reduction of glomeruli in the AL of honeybee drones is higher in the mALT- compared to the lALT-innervated hemilobe of the AL. ORN axons in the glomeruli of the T3 cluster are innervated by mALT PNs in females (Kirschner et al. 2006). This cluster is reduced in drones, whereas the T1 cluster is less reduced compared with that in the female AL. As drones lack S. basiconica, the reduction of the mALT-associated parts of the T3 glomeruli is likely to be related to the absence of ORNs from the S. basiconica. As three enlarged glomeruli (macroglomeruli, as indicated in Fig. 7) are present in the T1 cluster in drones (Sandoz 2006; Nishino et al. 2009), the slight reduction in the number of T1 glomeruli in drones might favor the macroglomeruli.
Several physiological studies, so far, have shown that no part of the honeybee AL and therefore neither the mALT or the lALT, is selectively specialized for either only social or floral odorants (Abel et al. 2001; Brill et al. 2013; Carcaud et al. 2012; Galizia et al. 2012; Krofczik et al. 2008; Müller et al. 2002; Rössler and Brill 2013; Yamagata et al. 2009). However, as honeybees are exposed to an enormous odor space in their natural environments, more odorants, in particular social (colony) cues and pheromones, remain to be tested in more detail and might give further indications concerning selective physiological properties and the molecular receptive range of S. basiconica ORNs. The ants Camponotus floridanus (Zube and Rössler 2008), Camponotus japonicus (Nishikawa et al. 2012) and A. vollenweideri (Kelber et al. 2010) have similarly been demonstrated to have a reduced number of glomeruli in males compared with females. Kelber et al. (2010) investigated projection patterns of S. trichoidea and S. basiconica ORNs in A. vollenweideri. Here, the ants possess six AL glomerular clusters (T1-T6) and S. trichoidea ORNs have been shown to project to all of them. Up to five different glomerular clusters have been found to be innervated by ORNs from a single S. trichoideum. In contrast, S. basiconica ORNs only projected to the T6 cluster (Kelber et al. 2010). In honeybee workers, we were able to reveal that S. basiconica preferentially project into the T3 cluster. C. japonicus ALs comprise seven glomerular clusters (T1-T7) and the glomerular cluster T6 is also worker-specific, similar to the T6 cluster in A. vollenweideri (Kelber et al. 2010; Nishikawa et al. 2012). Furthermore, here the T6 output neurons were demonstrated to project to specific subregions within the MBs and the LH and the authors speculate that the input to these glomeruli origins from S. basiconica-associated ORNs (Nishikawa et al. 2012). Nishikawa et al. (2012) argued that these brain regions are likely to be involved in social tasks in C. japonicus, as drones do not need to fulfil extensive social duties within the colony.
Another important aspect is that in several solitary bee species, the males also lack S. basiconica (Ågren 1977, 1978; Ågren and Hallberg 1996; Galvani 2012; Streinzer et al. 2013). In parasitoid wasps, cuticular hydrocarbon profiles have been shown to vary between closely related species, between sexes and according to the developmental environment (Khidr et al. 2013). Better detection abilities for cuticular hydrocarbons could serve females in kin detection and thus help them to avoid insemination by related males or even by males from other species. Another major distinction between drones and workers in both ants and social bees is that drones do not forage. This implies the involvement of S. basiconica in the detection of floral odorants. These studies show that various hymenopteran species possess female-specific olfactory subsystems consisting in specific sensilla, their respective ORNs and downstream odor-processing brain structures. The separation of these subsystems from the sexual isomorphic structures can be more (A. vollenweideri, C. japonicus) or less (A. mellifera) pronounced. Thus, variations in olfactory tasks and structures between males and females might differ across species.
Based on these differences described so far with regard to behaviors between male and female honeybees, we conclude that the S. basiconica might equally well be specialized for flower odorant detection or the detection of social olfactory cues or might even be generalistic.
Honeybee workers possess a specific olfactory subsystem comprising S. basiconica, parts of the T3 cluster of glomeruli and a significant proportion of mALT PNs. Drones completely lack this olfactory subsystem (Fig. 7) and, additionally, have far fewer S. trichoidea, all features that might favor a more elaborated queen-pheromone-processing system and the associated higher numbers of S. placodea present in drones (Brockmann et al. 2006; Kaissling and Renner 1968; Sandoz 2006). At the behavioral level, drones, therefore, are likely to have more limited odor discrimination and recognition abilities compared with females. This limitation is likely to be associated with social (colony) odors and/or floral odors. The adaptation in drones for improved queen-pheromone detection including high numbers of S. placodea and pheromone-processing macroglomeruli is likely to increase mating probabilities. Honeybee workers, in contrast, are exposed to high selective pressure to identify and locate correctly a wide variety of odorants, including floral odorant mixtures, a large variety of pheromones and colony (social) cues. The different types of sensilla and the associated olfactory subsystems of glomeruli and output tracts in the AL appear to be well adapted for these tasks.
We thank Cornelia Grübel for expert help with tract staining and confocal microscopy, Dirk Ahrens for beekeeping and Johannes Spaethe for fruitful discussions and help with statistics.
- Ito K, Shinomiya K, Ito M, Armstrong JD, Boyan G, Hartenstein V, Harzsch S, Heisenberg M, Homberg U, Jenett A, Keshishian H, Restifo LL, Rössler W, Simpson JH, Strausfeld NJ, Strauss R, Vosshall LB, Insect Brain Name Working Group (2014) A systematic nomenclature for the insect brain. Neuron 81:755–765PubMedCrossRefGoogle Scholar
- Mobbs PG (1982) The brain of the honeybee Apis mellifera. I. The connections and spatial organization of the mushroom bodies. Proc R Soc Lond [Biol] 298:309–354Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.