Thymol Affects Congruency Between Olfactory and Gustatory Stimuli in Bees

Abstract

Honey bees learn to associate sugars with odorants in controlled laboratory conditions and during foraging. The memory of these associations can be impaired after exposure to contaminants such as pesticides. The sub-lethal effects of acaricides such as 5-methyl-2-(propan-2-yl)-phenol (thymol) introduced into colonies to control varroa mites are of particular concern to beekeeping, due to detrimental effects of some acaricides on bees. Here we assess whether various odorant/sugar pairs are identically memorized in a differential appetitive olfactory conditioning experiment and whether this learning is affected by thymol exposure. Responses to odorants in retrieval tests varied according to the sugar they were paired with, a property called congruency. Interestingly, congruency was altered by pre-exposure to some thymol concentrations during retrieval tests, although electroantennography recordings showed it left odorant detection intact. This highlights the importance of taking into account subtle effects such as odor/sugar congruency in the study of the effect of pesticides on non-target insects, in addition to the simpler question of memory impairment.

Figure 1

Effect of thymol on the bee olfaction. Electroantennogram responses in mV (mean with standard error) for increasing concentrations of 1-hexanol, 2-hexanol, 1-octanol and 2-octanol diluted in mineral oil after exposure to thymol. Each curve corresponds to a pre-exposure to a thymol concentration; values in parenthesis are the sample size (the same animals were used but some were not tested with 2-octanol).

Figure 2

Bee PER rate during the learning trials. Each curve corresponds to the response to the CS+ or the CS− for different thymol treatments. As there was neither significant difference between the different odorants and sugars, nor any interactions, the odorant and sugar groups were pooled. A total of 725 bees were used (20–42 in each combination odorant/sugar/treatment, see Table 1). The initial high response rate to CS− is a usual observation in differential conditioning and corresponds to generalization from CS+.

Figure 3

PER rates for the CS+ during the retrieval tests performed 1 hour and 24 hours after the training reported in Fig. 2 (the corresponding CS− data are in Fig. 4). The upper radar plot shows scores of bees trained with fructose the lower plot bees trained with sucrose; plots on the left are for 1-hour retrieval tests, and plots on the right for 24-hour retrieval tests. The curves correspond to the different thymol treatment. The tip of the plots corresponds to the odorants used as CS+. Stars denote significant interaction between 1 µg thymol treatment, fructose and 1-octanol (logistic regression, **p < 0.010) or 1 µg thymol and 1-octanol (logistic regression, *p < 0.050); this kind of interaction is the hallmark of congruency alteration.

Figure 4

PER rates for the CS− during the retrieval tests performed 1 hour and 24 hours after the training reported in Fig. 2 (the corresponding CS+ data are in Fig. 3). Star denotes a significantly lower PER rate in response to 2-hexanol, irrespective of the sugar used or the thymol treatment (logistic regression, *p < 0.050). Other details are as in Fig. 3.