Data from: Social cues on stone tools outweigh raw material properties in wild primates
Data files
Jun 16, 2026 version files 4.71 MB
-
all.tasks.csv
441.65 KB
-
data.task2.site.csv
66.27 KB
-
README.md
6.04 KB
-
statistics_Rscript.r
43.08 KB
-
statistics_Rworkspace.RData
4.15 MB
Abstract
The dataset contains the data and statistical analyses for the manuscript “Social cues on stone tools outweigh raw material properties in wild primates”. In this manuscript, we performed three tasks to investigate bearded capuchin monkeys' (Sapajus libidinosus) ability to discriminate between stones of different hardness for nut-cracking. We tested 12 wild-living individuals at Serra de Capivara National Park (SCNP), and 16 semi-wild individuals at the Park Ecológico do Tietê (PET), Brazil. All monkeys were individually recognized and their age known from ongoing long-term studies.
In Task 1, capuchin monkeys were exposed to four stones (two hard sandstones and two softer limestones). In Task 2, they were exposed to two nut-cracking sites, each containing one hard sandstone and one hard limestone. One of the sites contained raw stones (as in Task 1), and the other contained modified stones and nutshells, indicating previous tool use. In Task 3, they were exposed to four stones, similarly to Task 1, but one of each stone type was raw and the other modified, indicating previous tool use.
For each trial, we recorded whether a stone was selected (1: yes, 0: no) as a tool (anvil or hammer), its stone ID, initial position in the setup (1st or 2nd left or right), as well as its material (hard sandstone or soft limestone) and percussive damage (raw stone or modified tool). For each focal individual participation, we noted the task number (1-3), the relative trial number within a task, the absolute trial number across tasks, as well as the individual characteristics (focal ID, sex, age class, park). To include the random effect of each focal individual's trial, we created a trialID with the individual's relative trial number (e.g., name. 1).
The data "all.tasks.csv" contains the described selections of the stones, as well as a note on why a trial was excluded (usually because a different object than the provided stones was used for nut-cracking).
The data "data.task2.site.csv" contains the selections of the tool sites (raw stones or nut-cracking), their positions (left or right), the presented stones, and which site was first chosen by the respective focal individual.
In the R script "statistics_Rscript.txt", the statistical models supporting the results of the manuscript are stored in the R workspace "statistics_Rworkspace.RData".
Dataset DOI: 10.5061/dryad.wm37pvn38
Description of the data and file structure
These data were collected as a part of the experimental design published in the paper entitled, "Social cues on stone tools outweigh raw material properties in wild primates". This paper was published in the Proceedings of the Royal Society B. The preprint can be found at https://doi.org/10.1101/2024.06.10.598233.
Files and variables
The following files can be used to reproduce the statistical analysis presented in the paper. The analysis is contained within a single self-contained R script named "statistics_Rscript.r." The data necessary to run the code can be accessed by loading the "statistics_Rworkspace" RData file in R. Once this is done, the script will run without issue. Copies of the dataset are also provided as csv files (semicolon-delimited), but it is not necessary to use these files when running the R code.
File: statistics_Rscript.r
Description: The self-contained script that runs all statistical analyses presented in the paper. For brevity, calculations for hammer selection are omitted since they followed the same structure as the calculations for anvil selection. It should be noted that some of the permutation tests are computationally intensive. Parallelization can be used to increase the computation speed, but the analysis can take days to run on personal laptops and desktops.
File: statistics_Rworkspace.RData
Description: The workspace containing all datasets, final model outputs, and bespoke functions necessary to run the script.
File: all.tasks.csv
Description: Contains the described selections of the stones, as well as a note on why a trial was excluded (usually because a different object than the provided stones was used for nut-cracking).
Variables
- ID: The unique identifier for the individual being observed.
- species: A categorical variable indicating the species of the participants.
- sex: A categorical variable indicating the sex of the participating individual (Male/Female).
- park: The specific park that the participating individual is a part of (SCNP/PET).
- age.class: A categorical variable indicating the general age of the individual. This is divided into 2 categories (adult, subadult).
- trialID: A unique identifier for the rel.trial per individual.
- abs.trial: The overall trial number as continuous value of participation.
- task: The task, either Task 1, Task 2, or Task 3.
- rel.trial: The trial number within the respective task.
- date: The date of data collection.
- time: The time of data collection.
- tool: The tool-category of the specific stone (anvil/hammer).
- stone.ID: A unique identifier for specific experimental stones, following the pattern "damage.materialNumber", e.g. r.L04 for raw limestone No. 4, or t.S22 for damaged (=tool) sandstone No. 22. Other objects used as anvil or hammer than the provided experimental stones are described (e.g., concrete, wood).
- position: The position of the specific stone in the setup, either on the right (R) or left (L), either on the outside (1) or the inside (2). If a different object than the offered experimental stones were used, the respective material is mentioned (e.g., wood, natural).
- material: The material of the stone (limestone/sandstone). If other objects were used than the provided stones, their material is described.
- selected: Identifier whether this stone.ID was used as this tool (1), or not (0).
- exclusion: Notion, why a trial was excluded.
- percussive: A categorical variable indicating whether a stone was raw, unused (raw), or modified by percussive damage (tool). Other objects used than the provided stones, are categorized as natural (stone) or wood.
File: data.task2.site.csv
Description: Contains the described selections of the tool sites (only Task 2).
Variables
- ID: The unique identifier for the individual being observed
- species: A categorical variable indicating the species of the participants.
- sex: A categorical variable indicating the sex of the participating individual (Male/Female).
- age.class: A categorical variable indicating the general age of the individual. This is divided into 2 categories (adult, subadult).
- park: The specific park that the participating individual is a part of (SCNP/PET).
- task: The task, either Task 1, Task 2, or Task 3.
- abs.trial: The overall trial number as continuous value of participation.
- rel.trial: The trial number within the respective task.
- date: The date of data collection.
- time: The time of data collection.
- left: The identification of the tool site on the left (raw stones/nut-cracking site)
- right: The identification of the tool site on the right (raw stones/nut-cracking site)
- L1: The unique stone.ID of the stone on the left side in the front.
- L2: The unique stone.ID of the stone on the left side in the back.
- R2: The unique stone.ID of the stone on the right side in the back.
- R1: The unique stone.ID of the stone on the right side in the front.
- site: The site which could be chosen (raw stones/nut-cracking site).
- sitechoice: Identifier that this site was chosen first (potential changes were not evaluated and are not in this table).
- side: The side of that site within the setup (left/right).
- trialID: A unique identifier for the rel.trial per individual.
- selected: Identifier whether this site was selected (1), or not (0).
Code/software
All calculations were done with R version 4.2.3 (R Core Team (2023). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/).
Access information
Other publicly accessible locations of the data:
- N/A
Data was derived from the following sources:
- N/A