Data and code for: Bee swimming is adaptive but disrupted by insecticide
Data files
Feb 17, 2026 version files 30.02 MB
-
Bee_swimming_paper_analysis.Rmd
16.70 KB
-
Fig1A.data.csv
921 B
-
Fig1B.data.csv
776 B
-
Fig1C.data.csv
111 B
-
Fig2A.data.csv
1.94 KB
-
Fig2B.data.csv
3.63 KB
-
Fig2C.data.csv
113 B
-
Fig2D.data.csv
4.58 KB
-
Fig2E.data.csv
808 B
-
Fig2F.data.csv
824 B
-
Fig3.data.csv
564 B
-
Fig4A.data.csv
2.29 KB
-
Fig4B.data.csv
2.21 KB
-
Fig4C.data.csv
2.40 KB
-
Fig4D.data.csv
137 B
-
Fig4E.data.csv
9.62 KB
-
Fig4F.data.csv
1.08 KB
-
Fig4G.data.csv
1.09 KB
-
README.md
7.96 KB
-
Supplementary_Movie_1.MOV
4.59 MB
-
Supplementary_Movie_2.MOV
9.65 MB
-
Supplementary_Movie_3.MOV
7.42 MB
-
Supplementary_Movie_4.MOV
3.02 MB
-
Supplementary_Movie_5.MOV
5.28 MB
-
TableS1.data.csv
135 B
Abstract
A unique type of locomotion was recently discovered that honey bees (Apis mellifera L.) could propel their own body on water surface by keeping their upper wing surface dry while moving their wings. However, it was not clear whether such locomotive behavior was ecologically meaningful. Here we show that honey bees preferred a dark region (skototaxis) while hydrofoiling on the water surface in trying to reach the edge. This was shown by the significant deviation from random distribution by Rayleigh tests and also by Contingency Table Analyses. However, this skototaxis was disrupted by a neonicotinoid insecticide thiamethoxam, as indicated by the random distribution of bees landing angle in pesticide-fed bees. This change is most likely due to a loss of motor control in the intoxicated honey bees, because they made 3 times more circular turns compared to the control bees. Finally, we show that mason bees (Osmia excavata) also displayed skototaxis, showing a significantly stronger preference for dark than honey bees. The female mason bees exhibited higher efficiency than males in swimming, as indicated by their shorter duration and distances. These findings suggest that swimming behavior in bees evolved before sociality and serves important adaptive and ecological functions. However, environmental pollution from excessive pesticide use may negatively impact this behavior.
Reuse of data is permitted. No honey bees were harmed during the experiments because all bees were returned to their colonies after the swimming tests were done. Data structure is described for each file in the readme file.
For methods, results and discussion, please see https://www.nature.com/articles/s42003-026-09669-w
Fig1A.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 85 honey bees as a control (they were fed with 50% sucrose water solution for 5 days, the sucrose contained 0.06% acetone to control for the insecticide treatment below).Experiment was performed in Michigan during Nov. 2019. These data were inputted to an R program to generate the circular plot in our paper (Fig. 1A).
Fig1B.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 68 honey bees that were fed with 125 ppb thiamethoxam inside 50% sucrose water solution for 5 days, the sucrose contained 0.06% acetone to dissolve the first stock solution of insecticide (not water soluble). Experiment was performed in Michigan during Nov. 2019. These data were inputted to an R program to generate the circular plot data in our paper (Fig. 1B).
Fig1C.data.csv
This file contains the number of bees landing in the dark area (piece of paper printed as black, 72 degrees wide) and in the white area (the rest of the circle, 80%). Expected is the 20% of landing in the dark area and 80% in the white area. Acetone.us is the bees treated with syrup containing small amount of acetone, and tmx.us is the treatment with thiamethoxam (125 ppb). These data were used to compute chi-square values in a Contingency Table Analysis. The results are presented in Fig. 1c.
Fig2A.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 98 honey bees as a control (they were fed with 50% sucrose water solution for 5 days, the sucrose contained 0.06% acetone to control for the insecticide treatment below). Experiment was performed in Guangzhou China in April 2024. These data were inputted to an R program to generate the circular plot data in our paper (Fig. 2A).
Fig2B.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 144 honey bees that were fed with 125 ppb thiamethoxam inside 50% sucrose water solution for 5 days, the sucrose contained 0.06% acetone to dissolve the first stock solution of insecticide (not water soluble). Experiment performed in Guangzhou China in April 2024. These data were inputted to an R program to generate the circular plot data in our paper (Fig. 2B).
Fig2C.data.csv
This file contains the number of bees landing in the dark area (piece of paper printed as black, 72 degrees wide) and in the white area (the rest of the circle, 80%). Expected is the 20% of landing in the dark area and 80% in the white area. Acetone.cn is the bees treated with syrup containing small amount of acetone, and tmx.cn is the treatment with thiamethoxam (125 ppb). These data were used to compute chi-square values in a Contingency Table Analysis. The results are presented in Fig. 2C.
Fig2D.data.csv
This file contains the swimming duration (starting being dropped onto water, to when bees land to the edge of glass bowl) of 242 honey bees, about half fed with sugar with some acetone and half fed with thiamethoxam (125 ppb) for 5 days. These were data from the China trial. These data were used to perform a t-test for their differences as presented in Fig. 2D.
Fig2E.data.csv
This file contains the swimming distance data (in cm), generated by EthoVision, from 20 randomly selected bees per group (acetone vs thiamethoxam). These were data from the China trial. These data were used to perform a t-test for their differences as presented in Fig. 2E.
Fig2F.data.csv
This file contains the swimming velocity data (cm/s) , generated by EthoVision, from 20 randomly selected bees per group (acetone vs thiamethoxam). These were data from the China trial. These data were used to perform a t-test for their differences as presented in Fig. 2F.
Fig3.data.csv
This file contains the number of turns made by each bee, from tracking data generated by EthoVision, from 20 randomly selected bees per group (acetone vs thiamethoxam). These were data from the China trial. These data were used to perform a t-test for their differences as
presented in Fig. 3.
Fig4A.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 106 honey bees as a control, these bees were sampled and tested the same day (after 2 hours inside a cage). Experiment performed in Guangzhou China during May 2024. These data were inputted to an R program to generate the circular plot data in our paper (Fig. 4A).
Fig4B.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 116 male mason bees. Experiment performed in Guangzhou China during March 2024. These data were inputted to an R program to generate the circular plot data in our paper (Fig. 4B).
Fig4C.data.csv
This file contains the landing position (angle measured from the left side boundary of the black stripe) of 121 female mason bees. Experiment performed in Guangzhou China during March 2024. These data were inputted to an R program to generate the circular plot data in our paper
(Fig. 4C).
Fig4D.data.csv
This file contains the number of bees landing in the dark area (piece of paper printed as black, 72 degrees wide) and in the white area (the rest of the circle, 80%). Expected is the 20% of landing in the dark area and 80% in the white area. Honeybee were Apis mellifera tested in Guangdong, China (as presented in Fig. 4A). Mason-m was male mason bees and mason-f was female mason bees (data presented in Fig. 4B, 4C). These data were used to compute chi-square values in a Contingency Table Analysis. The results are presented in Fig. 4D.
Fig4E.data.csv
This file contains data of honey bees, male and female mason bees swimming duration (seconds). These were based on the video files and not extracted data from EthoVision. The data were used to conduct ANOVA and presented in Fig. 4E.
Fig4F.data.csv
This file contains data of honey bees, male and female mason bees swimming distance (cm). These data were provided by EthoVision after generating the bees swimming trajectories. The data were used to conduct ANOVA and presented in Fig. 4F.
Fig4G.data.csv
This file contains data of honey bees, male and female mason bees swimming velocity (cm/s). These data were provided by EthoVision after generating the bees swimming trajectories. The data were used to conduct ANOVA and presented in Fig. 4G.
TableS1.data.csv
This file contains data of the honey bee three trials (honeybee1 for Michigan trial, honeybee2 for China insecticide trial, and honeybee3 was the one trial together with the mason bees). These data were used to show that the three trials were very consistent with 42 to 49% of bees landing in the 20% black area. These data were not significantly different from one another by Contingency Table Analysis.
Bee_swimming_paper_analysis.Rmd
This file contains the R codes each of which reads the data files (.csv files) listed above and also their outputs.
Supplementary_Movie_1.MOV
A honey bee worker fed with sugar (containing 0.06% acetone) swam on the water surface and landed in the dark region.
Supplementary_Movie_2.MOV
A honey bee worker fed with thiamethoxam swam on the water surface, making numerous turns before landing in a region other than the dark area.
Supplementary_Movie_3.MOV
A male mason bee swam on the water surface and landed in the dark region.
Supplementary_Movie_4.MOV
A female mason bee swam on the water surface and landed in the dark region.
Supplementary_Movie_5.MOV
A honey bee worker fed only with sugar swam on the water surface and landed in the dark region.
Detailed methods are described in the published paper.
