In-hive learning of specific mimic odours as a tool to enhance honey bee foraging and pollination activities in pear and apple crops

Abstract

The areas devoted to agriculture that depend on pollinators have been sharply increased in the last decades with a concomitant growing global demand for pollination services. This forces to consider new strategies in pollinators' management to improve their efficiency. To promote a precision pollination towards a specific crop, we developed two simple synthetic odorant mixtures that honey bees generalized with their respective natural floral scents of the crop. We chose two commercial crops for fruit production that often coexist in agricultural settings, the apple (Malus domesticus) and the pear trees (Pyrus communis). Feeding colonies with sucrose solution scented with the apple mimic (AM) or the pear mimic (PM) odour enabled the establishment of olfactory memories that can bias bees towards the flowers of these trees. Encompassing different experimental approaches, our results support the offering of scented food to improve foraging and pollination activities of honey bees. The circulation of AM-scented sucrose solution inside the hive promoted higher colony activity, probably associated with greater activity of nectar foragers. The offering of PM-scented sucrose solution did not increase colony activity but led to greater pollen collection, which is consistent with pear flowers offering mainly pollen as resources for the bees. Results obtained from apple and pear crops suggest that the offering of AM- and PM-scented sucrose solution increased fruit yields. This preliminary study highlights the role of in-hive olfactory learning to bias foraging preferences within pome fruit crops.

Figure 1

Generalization of memories from pear mimic odours to natural floral scents. (A) Generalization was tested towards the single presentation of the pear natural odour (right panel) after one of the pear mimics (PM, PMI or PMII) was used as conditioned stimulus (CS) during an absolute classical conditioning of the proboscis extension reflex (PER; left panel). Asterisks indicate significant differences between responses obtained at the third conditioning trial and the test (***, p < 0.001). No significant difference (n.s.) indicates that bees could not discriminate between PM (3rd trial) and the unrewarded pear natural scent (PER-test). (B) Generalization was evaluated towards the single presentation of the natural apple and pear odour (right panel) after PM (from A) was used as CS during the absolute PER conditioning (left panel). Same letter indicates no significant differences between the generalized response of pear natural odour and the response retrieved by the PM itself. The experimental subjects were all foraging bees completely naïve for the CS that had no access to any pear tree. Symbols indicate the proportion of PER and bars (in test) show the 95% confidence intervals. Numbers between brackets indicate sample size. Package ‘lme4’, version 1.1-30. https://github.com/lme4/lme4/.

Figure 2

Discrimination between mimic odours and natural floral scents. (A,B) Discrimination was evaluated towards the single presentation of the apple natural floral odour and the jasmine mimic (JM) (right panels) after a differential proboscis extension reflex (PER) conditioning (left panel), where both apple natural odour and JM were used as rewarded (CS+) and non-rewarded stimulus (CS−). Asterisks indicate significant differences between tested odours (***, p < 0.001). (C,D) Differential PER conditioning between the apple natural floral odour and the apple mimic (PM), where both odours were used as rewarded (CS+) and non-rewarded stimulus (CS−). No difference (n.s.) at test indicates that bees could not discriminate between AM and the unrewarded apple natural scent (D). (E,F) Differential PER conditioning between the pear natural floral odour and the apple mimic (AM), where both odours were used as rewarded (CS+) and non-rewarded stimulus (CS−). Asterisks indicate significant differences between tested odours (***, p < 0.001) showing that bees could discriminate between AM and the pear natural scent. The experimental subjects were all foraging bees naïve for the CSs that had no access to any pear or apple tree. Numbers between brackets indicate sample size. Circles indicate the proportion of PER and bars (in test) show the 95% confidence intervals. Package ‘emmeans’, version 1.8.0. https://github.com/rvlenth/emmeans.

Figure 3

Effect of the mimic odours on the activity of colonies that pollinate apple or pear trees. Percentage of blooming of the dominant variety of apple tree (A) or pear trees (B) during the experimental period. Colonies were fed either unscented sucrose solution (SS) or apple mimic-scented sucrose solution (SS + AM) (C) or pear mimic-scented sucrose solution (SS + PM) (D). Incoming bees were calculated as the counts after the offering of the solutions (2, 3, 4, 5, 6, 7 and 9 days). Asterisks indicate significant differences between the specified treatments (***, p < 0.001; **, p < 0.01). Circles indicate the mean values and bars show the 95% confidence intervals. Numbers between brackets indicate the number of colonies monitored per treatment. Package ‘glmmTMB’ version 1.1.4. https://github.com/glmmTMB/glmmTMB.

Figure 4

Effect of the pear mimic odour on the pollen foraging. (A) Percentage of blooming of the dominant variety of pear tree during the experimental period. As an example is shown one colony fed unscented sucrose solution (SS) and one colony pear mimic-scented sucrose solution (SS + PM). (B) Size of pear pollen loads trapped at the colony entrance is expressed in mg of 10 pollen loads along the experimental period. (C) Number of pollen loads accumulated during days 3, 7 and 8 after the offering of pear mimic-scented sucrose solution (SS + PM) or unscented sucrose solution (SS) for the same colonies. Pollen traps were placed in one colony treated of each treatment, the samples obtained were classified according to their colour into the categories: Pyrus communis (light grey bar) and pollen from other species (dark grey bar).

Figure 5

Effect of the offering of mimic-scented food on fruit yield and on plantation yield. (A,B) Yield was calculated as the counts of fruits per trees in a pair of apple plots and in a pair of pear plots, where 30 trees were surveyed in each. (A) Colonies that pollinate each apple plot were fed apple mimic-scented sucrose solution (SS + AM) or unscented sucrose solution (SS). (B) Colonies that pollinate each pear plot were fed pear mimic-scented sucrose solution (SS + PM) or unscented sucrose solution (SS). (C,D). Yield was obtained either from 11 apple plots (varieties: Hi Early, Granny Smith and Chañar 28) that settle 130 colonies in total that had been fed apple mimic-scented sucrose solution (SS + AM) and from 11 apple plots (varieties: Hi Early, Granny Smith and Chañar 28) with 139 colonies that had been fed unscented sucrose solution (SS). Yields are presented for all apple varieties together (C) or separately (D) Asterisks indicate significant differences between the treatments (*, p < 0.05; **, p < 0.01; ^#, p = 0.06). Circles indicate the mean values and bars show the 95% confidence intervals. Package ‘ml’, version. Package ‘emmeans’, version 1.8.0. https://github.com/rvlenth/emmeans.