We used five M. beecheii (Hymenoptera, Apidae, Meliponini) colonies obtained from a meliponary in Tuxtla Chico (14°56'25"N, 92°10'08"W), Chiapas, Mexico. The experiments were conducted from April to October 2021 at El Colegio de la Frontera Sur gardens, Tapachula (14°54'39.86"N, 92°15'51.55"W), Chiapas, Mexico. Annual rainfall in this region is approximately 3843 mm, with September as the wettest month and February as the driest month. The temperature normally ranges between 29 °C and 35 °C. The colonies were free of fungi and parasites.

The forager bees were trained to collect a solution of 3 M sucrose ad libitum from an artificial feeder, consisting of a Petri dish (100 × 10 mm) containing a small cotton ball drenched in the sucrose solution in the center. The feeder was placed 5 m from the beehives. The training was conducted between 08:00 and 13:00 h.

We extracted the compounds deposited by M. beecheii foragers on a glass feeder (100 × 10 mm) (Hrncir et al. 2004) by washing the feeders with 4 mL hexane (HPLC grade, Aldrich, Toluca, Mexico). The amount of compounds left by 40 or 50 foragers on the food source was considered as a biological active mark. The hexane extracts were concentrated to 50 μL using a gentle flow of N₂ and stored at -20 °C until their analysis. Five different extracts were obtained.

To prepare the gland extracts, we captured foragers bees that arrived at the feeder during training. Bees were frozen at -20 °C before dissection and analysis. The glands of the foragers were carefully dissected in distilled water with two pairs of fine tweezers under a stereoscopic microscope. The gland extracts were prepared by carefully dissecting the labial and mandibular glands of 10 foragers in 1000 μL of solution. So, 100 μL of the prepared solution is the amount corresponding to 1 labial gland equivalent (LGE), 50 μL of solution corresponds to 0.5 LGE and 10 μL to 0.1 LGE; and 100 μL of the prepared solution is the amount corresponding to 1 mandibular gland equivalent (MGE), 50 μL of solution corresponds to 0.5 MGE and 10 μL to 0.1 MGE. Five gland extracts were prepared.

In a first experiment, we evaluated whether the forager bees leave odor marks on the feeders in a two-choice tests. We offered to foragers two feeders: one with chemical marks (previously visited by foragers) and the other clean (not visited by foragers). The feeders were placed at the site where the bees were trained. Feeders were placed 30 cm from each other. We placed a few drops of the 3M sucrose solution in the entrance to stimulate the visits. We recorded the number of bees that visited the marked feeder and the clean feeder. A visit was counted when the bee landed and extended its proboscis to feed. All bees were marked with a fine brush with acrylic paint not toxic and captured on their first visit to avoid counting the same bee more than once (Alavez-Rosas et al. 2017). To prevent the phenomenon of social facilitation (influence in the bee election due to the presence of a conspecific at the feeding site), care was taken not to count the bees that visited the feeders while other bees were there. The position of the feeders was interchanged every 5 min to prevent position bias. At the end of the trial, the bees were freed near the colony. All trials were conducted between 08:00 and 13:00 h. In a total, 10 replications were performed.

In a second experiment, we assessed the effect of the labial gland extracts on food searching in two choice tests. Here, using pieces of 1cm² filter paper placed on feeders, one feeder was sprinkled with 10 μL of labial gland extract at the beginning of the experiment and the other feeder was sprinkled with 10 μL of hexane as a control. The gland extracts were evaluated at 0.1, 0.5, and 1 LGE. In total, 10 replications were performed for each gland equivalent extract.

In a third experiment, we evaluated foragers preference for labial gland extract, mandibular gland extract, or solvent in three-choice tests. Three feeders were placed in the training site: control (hexane), labial gland extract, and mandibular gland extract. The feeders were placed 20 cm from each other. Care was taken not to count bees that visited the feeder while other bees were there to avoid the phenomenon of social facilitation. The position of the feeders was interchanged every 5 min to avoid position bias. The gland extracts were evaluated at 0.1, 0.5, and 1 LGE or MGE. In total, 10 replications were performed for each gland equivalent extract.

We collected forager bees from three established colonies, bees were frozen for one minute to numb them before dissection, subsequently their antennae were carefully removed. The base of the antenna was inserted into the tip of the glass capillary logging electrode. The signals generated by the antenna passed through a high impedance amplifier (NL 1200; Syntech, GmbH) and visualized with the software Syntech to process the EAG signals. We used a stimulus Flow controller (CS-05, Syntech) to generate stimuli at intervals of one minute. A constant current of pure humidified air (0.7 L min^-1) was directed toward the antenna (Malo et al. 2004).

The experimental procedure consisted of depositing the treatment (1 LGE extract, 1 MGE extracts or solvent, in this order) onto 1.5 × 1.5 mm pieces of filter paper (Whatman no. 1, Whatman, Maidstone, England) exposed to air for 20 s to allow the solvent to evaporate, inserted into a glass Pasteur pipette, and left for 40 s before applying. The application of the stimulus consisted of inserting the tip of the pipette that contained the piece of filter paper in a hole at the end of the glass tube through which the current of air blew continuously on the antenna. The waiting time between one stimulus and the next was one minute. The treatment was carried from the filter paper to the antenna on the controlled air current (0.5 L/min). The duration of the stimulus was 1 s. The continuous flow of pure air was maintained through the tube to assure that the odors were removed immediately. We used one antenna of the bee per treatment, and at least thirty five bees were used.

Extracts were analyzed in a gas chromatograph (Shimadzu GC-2010 Plus) coupled with a quadrupole mass spectrometer (Shimadzu, TQ8040), using a capillary column of non-polar silica SPB-1 (30 m long × 0.25 mm interior diameter) (Supelco, Toluca, Mexico). The initial temperature was 50 °C (held for 2 min), increased 15 °C/min up to 280 °C (held 10 min). Helium was the carrier gas and the injector temperature was 250 °C. Ionization was achieved by electron impact at 70 eV. The compounds were identified by comparison with those registered in the NIST 2014 library (software GCMS-solution), the retention index, the mass spectra, and retention times of available synthetic standards. The relative percentage of the components was calculated from the sum of the recorded peaks.

We evaluated the mixture of synthetic compounds with some compounds identified in the labial gland secretion. The synthetic blend evaluated was prepared in accord with the natural proportions of the M. beecheii labial gland using hexane as solvent. The compounds evaluated were heneicosane (100 ng), tricosane (100 ng), and pentacosane (25 ng). These compounds were chosen as they were available in supplies. Unfortunately, C27:1 isomers the main components in the labial gland secretion were not commercially available. We recorded the number of bees that visited a feeder baited with the synthetic blend, or a feeder with solvent. The feeders were placed where the bees were trained to visit. The distance between the feeders was 30 cm. The position of the feeders was changed every 5 min to avoid position bias. In total, 10 replications were performed in this experiment.

All data were analyzed using R software (R Core Team 2020). The EAG data were natural logarithm transformed and analyzed by a one-way analysis of variance (ANOVA), followed by a Tukey test. Behavioral data were transformed to satisfy the assumptions of normality and homoscedasticity and analyzed by generalized linear model (GLM) to the Poisson or negative binomial models.

M. beecheii foragers preferred the feeders visited previously by their conspecifics over the clean feeder (χ2 = 56.783, df = 1, p < 0.001) (Fig. 1).

Figure 1. 

Mean (± SEM) responses of M. beecheii foragers to marked and clean feeders (*** P< 0.001). Ten replications were performed.

In the two-choice bioassays, foragers preferred to visit the feeders marked with extracts of 0.5 LGE (χ2 = 134.38, df = 1, p < 0.001) and 1 LGE (χ2 = 71.676, df = 1, p < 0.001) over those treated with hexane. However, they did not show a preference for the 0.1 LGE or the control (χ2 = 0.53236, df = 1, p = 0.4656) (Fig. 2).

Figure 2. 

Mean (± SEM) responses of foraging M. beecheii worker bees to labial gland extract from conspecific worker bees at different concentrations (LGE=labial gland equivalent, *** P< 0.001, ** P<0.01, * P< 0.05). Ten replications were carried out for each gland equivalent extract.

In the three-choice bioassays, more foragers preferred visiting feeders with labial gland extracts (0.1 LGE: χ2= 13.335, df = 2, p<0.01; 0.5 LGE: χ2= 81.747, df = 2, p < 0.001; 1 LGE: χ2 = 23.929, df = 2, p < 0.001) over other feeders with mandibular extracts or the control (Fig. 3).

Figure 3. 

Mean (± SEM) responses of M. beecheii foragers to labial gland and mandibular gland extracts at different concentrations. Different small letters (P < 0.05) indicate significant differences between treatments. Ten replications were performed for each gland equivalent extract. LGE=labial gland equivalent, MGE=mandibular gland equivalent, C=control.

Analysis of EAG data revealed significant differences in the antennal response of forager bees to the different extracts evaluated and to the solvent (control) (F = 13.24, df = 2, P > 0.001). The antennal responses of the foragers was greater with 1 eq/µL of labial gland extract than with the mandibular gland extract or with the control (Fig. 4).

Figure 4. 

Electroantennographic (EAG) response in mV of M. beecheii forager bee antennae to labial gland extract, mandibular gland extract and the control (C). Different small letters (P < 0.05) indicate significant differences between treatments. N=35. LGE=labial gland equivalent, MGE=mandibular gland equivalent, C=control.

Chemical analysis showed that the labial gland extracts and extracts from feeders visited by foragers contain a mixture of unsaturated hydrocarbons, followed by straight chain hydrocarbons and small quantities of esters. The main components are a mixture of alkene isomers C27:1 (Table 1).

The chromatographic profile of the hexane feeder wash was similar to the labial gland extract profile, but different from that of the mandibular gland extract (Fig. 5).

Figure 5. 

Chromatographic analysis of the extracts from M. beecheii A hexane wash of the Petri dish feeder B hexane extract of the labial gland C hexane extract of the mandibular gland.

Foragers did not show a preference for the feeders treated with a three-component blend and those treated with hexane.