Figure 1.
Proportion of non-contaminated flower visitation between experiments.
The bars represent the means between the different colonies and their 95% confidence interval. The foragers were feeding more often on the non-contaminated flower when the other one was contaminated by a pathogen. This proportion increased when the other flower was contaminated with C. bombi (GLMM: p<0.001; C. bombi vs control: p<0.001, C. bombi vs E. coli: p<0.001, E. coli vs control: p<0.01).
Figure 2.
Feeding duration, flower preference and proportion of uncontaminated flower visitation for E. coli experiment.
A) Feeding duration on both flowers with and without the presence of Escherichia coli (n = 1150), B) Visit duration on both flowers with and without the presence of Escherichia coli (n = 1150), C) Proportion of non-contaminated flower visitation for E. coli experiment. C (in white) represents the presence of the parasite in the flower and NC (in grey) its absence. For the feeding duration, box plots depict median, interquatile range and non-outlier range; the dots represent the outliers. The bars represent the means between the different colonies and their 95% confidence interval. Foragers feed longer on the uncontaminated flower (GLMM: p<0.05), visit it more often (GLMM: p<0.01) and are more accurate when the flower is on left position (GLMM: p<0.001).
Figure 3.
Feeding duration, flower preference and proportion of uncontaminated flower visitation for C. bombi experiment.
A) Feeding duration on both flowers with and without the presence of Escherichia coli (n = 1400), B) Visit duration on both flowers with and without the presence of Escherichia coli (n = 1400), C) Proportion of non-contaminated flower visitation over days and between sympatric population (grey dot & continuous line) and allopatric population (black triangle & dashed line) for C. bombi experiment. C (in white) represents the presence of the parasite in the flower and NC (in grey) its absence. For the feeding duration, box plots depict median, interquatile range and non-outlier range; the dots represent the outliers. The bars represent the means between the different colonies and their 95% confidence interval. Foragers spend the same time feeding on both flowers (GLMM: p = 0.24), visit preferentially the uncontaminated flower (GLMM: p<0.001). The proportion of uncontaminated flower visitation increase over days and for the sympatric population this increase is stronger than for the allopatric population (GLMM: p<0.01; factor day: p<0.05, interaction between day and population's origin: p<0.01).
Figure 4.
Proportion of visits on the flower without Crithidia bombi over days between the two groups of experienced foragers: frequent (n = 10) and rare flyers (n = 26).
The black triangles and dash line represent the frequent flyers group and the grey squares and continuous line the rare flyers group. The symbol represent the mean and the bars the standard error. On the first day, the frequent flyers visited more often the flower without parasite than the rare flyers (M-W-U-test: Z = −2.40, p<0.05) but over days the rare flyers increased their proportion of visits on the flower where the parasite is absent to reach the same level than the frequent flyers (Friedman ANOVA: χ2 = 9.15, p<0.01; 2nd day: M-W-U-test: Z = 0.77, p = 0.45; 3rd day: M-W-U-test: Z = 1.49, p = 0.15).