Group phenotypic composition drives task performances in ants
Data files
Dec 20, 2023 version files 39.53 KB
-
Data_Martin_et_al.xlsx
-
README.md
Abstract
Differences in individual behaviour within a group can give rise to functional dissimilarities between groups, particularly in social animals. However, how individual behavioural phenotypes translate into the group phenotype remains unclear. Here we investigate whether individual behavioural type affects group performance in an eusocial species, the ant Aphaenogaster senilis. We measured individual behavioural traits and created groups of workers with similar behavioural type, either high-exploratory or low-exploratory. We tested these groups in four different, ecologically relevant tasks: reaction to an intruder, prey retrieval from a maze, nest relocation and tool use. We show that, compared to groups of low-exploratory workers, groups of high-exploratory workers were more aggressive towards intruders, more efficient in collecting prey, faster in nest relocation and more likely to perform tool use. Our results demonstrate a strong link between individual and collective behaviour in ants. This supports the ‘behavioural type hypothesis’ for group dynamics, which suggests that an individual’s behaviour in a social environment reflects its own behavioural type. The average behavioural phenotype of a group can therefore be predicted from the behavioural types of individual group members.
README: Group phenotypic composition drives task performances in ants
https://doi.org/10.5061/dryad.31zcrjdsm
The data were collected at the Laboratory of Experimental and Comparative Ethology, University Sorbonne Paris Nord, using 8 different colonies of the ant species Aphaenogaster senilis for the repeatability tests and 6 different colonies for the group tests.
The methodology used for data collection is described in the Manuscript entitled “Group phenotypic composition drives task performances in ants” submitted to Biology Letters.
1. Repeatability (Boldness) data sheet:
Contains the variable (Ex_lat) collected during the Boldness test to assess the repeatability of behavioural traits. This variable was collected with a chronometer and written down in a spreadsheet (.xls).
Variables included:
· Ex_lat: latency to exit the tube (sec)
· ID: worker IDs
· Nr_of_session: which one of the two sessions of the Boldness test
2. Repeatability (Openfield) data sheet:
Contains the variables (Int and ExtM) collected during the Open-field tests to assess the repeatability of behavioural traits.
Variables included:
· Int: percentage of time the individual spent inside the central area
· ExtM: percentage of time the individual spent moving in the periphery
· ID: worker IDs
· Nr_of_session: which one of the two sessions of the Open-field test
Int and & ExtM variables were extracted from the video-recordings using a custom-made ant tracking software for open-field test. When the ant moves it creates a difference between two consecutives frames of the video enabling us to track her movement (more details in supplementary material). The software then calculates the percentage of frames the individual spent moving in the periphery (directly equivalent to ExtM) and the percentage of frames the individual spent in the central area (directly equivalent to Int). A csv file with the frame percentages was given as the output of the software.
The following variables were collected during the Group behaviour assays. The variables used in the analysis of these assays are highlighted in the Manuscript with the symbols “i-xii”, also shown in the following description of the variables in parentheses.
3. Tool use data sheet:
Contains the variables (FirstTool, FirstHoney, UsingTools) collected during the Tool Use assay to assess the differences in the behaviour between the two types of sub-colonies. The experiments were filmed (.mp4), the variables were collected by analysing the video recordings by hand and written down in a spreadsheet (.xls).
Variables included:
· FirstTool (i): latency of first interaction with a tool (sec)
· FirstHoney (ii): latency of first interaction with the honey (sec)
· UsingTools: sub-colony performing tool use (1) or not (0)
· Subcolony: high- (High) and low–exploratory (Low) sub-colonies
· Col: colony IDs
4. Reaction to Intruder data sheet:
Contains the variables (Anten, MandOpen, Bite, Tot) collected during the Reaction to Intruder assay to assess the differences in the behaviour between the two types of sub-colonies. The experiments were filmed (.mp4), the variables were collected by analysing the video recordings by hand and written down in a spreadsheet (.xls).
Variables included:
· Anten (iii): proportion of time spent with intruder antennation (sec)
· MandOpen (iv): proportion of time spent with mandible opening towards the intruder (sec)
· Bite (v): Proportion of time spent with biting the intruder (sec)
· Tot: total time of interaction with the intruder (sec)
· Subcolony: high- (High) and low–exploratory (Low) sub-colonies
· Col: colony IDs
5. Prey retrieval from a maze data sheet:
Contains the variables (FirstEnter, FirstCont, FirsttoNest, NoDroso) collected during the Prey retrieval from a maze assay to assess the differences in the behaviour between the two types of sub-colonies. The experiments were filmed (.mp4), the variables were collected by analysing the video recordings by hand and written down in a spreadsheet (.xls).
Variables included:
· FirstEnter (vi): latency of first entrance into maze (sec)
· FirstCont (vii): latency of first contact with a Drosophila (sec)
· FirsttoNest (viii): latency to transport the first Drosophila to nest (sec)
· NoDroso (ix): number of Drosophila corpses transported into nest
· Subcolony: high- (High) and low–exploratory (Low) sub-colonies
· Col: colony IDs
6. Nest relocation data sheet:
Contains the variables (DiscNest, FirstLarvaToNest, NoWork) collected during the Nest relocation assay to assess the differences in the behaviour between the two types of sub-colonies. The experiments were filmed (.mp4), the variables were collected by analysing the video recordings by hand and written down in a spreadsheet (.xls).
Variables included:
· DiscNest (x): latency to discover new nest site (sec)
· FirstLarvaToNest (xi): latency to transport first brood to nest (sec)
· NoWork (xii): number of workers involved in nest relocation (carrying brood)
· Subcolony: high- (High) and low–exploratory (Low) sub-colonies
· Col: colony IDs
The variables were used in the statistical analyses as described in the Supplemental information accompanying the Manuscript entitled “Group phenotypic composition drives task performances in ants”. All statistical analyses were performed in R Statistical Environment, version 4.2.2.
Code/Software
Tracking and data processing
The tracking of the individual ants was performed by a program written in Python using the OpenCV computer vision package. The source code is available on the following GitHub page: https://github.com/Rayanne-M/Openfield-Ant-Tracking. The tracking is based on the movement of the ant, by differentiating between two consecutive frames. The general principle of the algorithm is based on the subtraction of these frames together and then filtered by a threshold. Thus, the position of the ant can be accurately determined, although it is sensitive to the change of background (a constant background is strongly recommended). The ant is considered to be standing if its position between two frames remains unchanged.