Data from: Understanding the caffeine paradox: caffeinated nectar increases pollinators’ perceived value of flowers
Data files
Mar 19, 2026 version files 697.37 KB
-
README.md
3.68 KB
-
stats_figs.html
655.58 KB
-
stats_figs.Rmd
13.04 KB
-
test.csv
5.61 KB
-
training.csv
19.47 KB
Abstract
Caffeine, nicotine, and cocaine are toxic plant defence compounds that are used to deter herbivory, yet they are rewarding, compulsively consumed, and addictive. This paradox remains poorly understood. The rewarding effects of plant-derived psychostimulants are thought to be an evolutionary side effect of their toxicity. However, caffeine and nicotine are found at low, nontoxic levels in plant nectars. Here, attention was directed towards caffeine, and whether caffeine changed pollinators’ valuation of flowers was examined. Bumble bees were trained to forage from two differently coloured flowers, with equal or different sugar concentrations separately on consecutive days or simultaneously. One flower type was caffeinated. Bees that experienced equal sugar concentrations from the two flower types showed no immediate preference, but they preferred the caffeinated flower after 24 h. In contrast, bees that experienced unequal sugar concentrations preferred the sweeter, noncaffeinated flower during training, but their preference for the caffeinated flower increased after 24 h compared with controls. These results indicate that caffeine influences bumble bee preferences by altering the perceived comparative value, rather than enhancing memorability. This finding provides a novel explanation for the paradox of plant-derived psychostimulants. That is, plants evolved to use compounds that target animal reward systems, thereby enhancing their perceived value to pollinators in comparison with other neighbouring flowers.
https://doi.org/10.5061/dryad.4b8gthtk7
The dataset contains all the raw data used for the statistical analyses and visualisation in the manuscript, (experiment 1-5). In each experiment, foraging options with non-caffeinated and/or caffeinated sugar water were provided. There are two data files, one for training and on for test: 'training.csv', 'test.csv'.
training.csv
In 'training.csv' there are seven columns:
- 'beeID' entries are the unique identification numbers for each individual bee used in each experiment;
- 'experiment' entries are the experiment conducted, with 'dif3030caf' = training bees with two options,30% w/w sugar water vs. 30% w/w + 0.13 mM caffeine sugar water simultaneously; 'dif4530caf' = training bees with 45% w/w sugar water vs. 30% w/w + 0.13 mM caffeine sugar water simultaneously; 'dif4530' = training bees with 45% w/w sugar water vs. 30% w/w sugar water, both options are non-caffeinated;
- 'colony' entries are the identification numbers for each colony;
- 'colour' entries are the counterbalanced colour sets used for different groups, with 1 = lime colour associated with the first option listed in the 'experiment' column, and organge colour associated with the second option listed in the 'experiment' column; 2 = vice versa;
- 'proportion' values are the proportion of choices for the second option listed in the 'experiment' column;
- 'block' entries are the numbers of the blocks in training that generate that values in the 'proportion' column;
- 'numchoice' values are the total number of landings in each black, which is identical here since each block contains 10 landings.
test.csv
In 'test.csv' there are seven columns:
- 'beeID' entries are the unique identification numbers for each individual bee used in each experiment;
- 'experiment' entries are the experiment conducted, with 'dif3030caf' = training bees with two options,30% w/w sugar water vs. 30% w/w + 0.13 mM caffeine sugar water simultaneously; 'dif4530caf' = training bees with 45% w/w sugar water vs. 30% w/w + 0.13 mM caffeine sugar water simultaneously; 'dif4530' = training bees with 45% w/w sugar water vs. 30% w/w sugar water, both options are non-caffeinated; 'abs3030caf' = training bees with 30% w/w sugar water on day 1 and 30% w/w + 0.13 mM caffeine sugar water on day 2; 'abs4530caf' = training bees with 45% w/w sugar water on day 1 and 30% w/w + 0.13 mM caffeine sugar water on day 2;
- 'colony' entries are the identification numbers for each colony;
- 'col_seq' entries are the counterbalanced colour sets used for differential training groups (dif3030caf, dif4530caf, dif4530), with 1 = lime colour associated with the first option listed in the 'experiment' column, and organge colour associated with the second option listed in the 'experiment' column; 2 = vice versa; those entries are the counterbalanced training sequence sets used for absolute training groups (abs3030caf, abs5430caf), with 1 = training with order listed in the 'experiment' column, and 2 = training with the reversed order;
- 'proportion' values are the proportion of choices for the second option listed in the 'experiment' column;
- 'numchoice' values are the total number of landings in the test for each bee;
- 'test' entries are the types of test: 'imme' = immediate test after training; '24h' = 24 hours after training.
Code/Software
stats_figs.Rmd is the R markdown file containing all statistical analyses and visualisation.
stats_figs.html is the knitted HTML output of stats_figs.Rmd, presenting full model summaries.