Data from: Lethal conflict after group fission in wild chimpanzees
Data files
Apr 01, 2026 version files 5.34 MB
-
0_threshold.zip
101.65 KB
-
77_focal_graph.zip
32.69 KB
-
chimp_behav_data.csv
4.91 MB
-
combined_membership_labels.csv
2.52 KB
-
learned_graphs_csv.zip
266.17 KB
-
patrol-data-quarterly.csv
2.71 KB
-
population_snapshots.csv
19.64 KB
-
README.md
4.60 KB
Abstract
Territorial conflicts in animals inform aspects of human warfare, but civil war, with its shifting group identities, appears to be uniquely human. We report a rare, permanent fission in the largest-known group of wild chimpanzees (Pan troglodytes). Using 30 years of behavioral observations and network analyses, we identify an abrupt transition from cohesion to polarization in 2015 and the emergence of two distinct groups by 2018. Over the next seven years, members of one of the splinter groups made 24 attacks, killing at least seven mature males and 17 infants in the other group. We provide the data and code used to analyze chimpanzee social networks. We provide the data and the code for other analyses, including figures on changes in group size, social networks, and mortality.
Dataset DOI: 10.5061/dryad.sf7m0cgkg
Description of the data and file structure
Behavioral observations on wild chimpanzees at Ngogo, Kibale National Park, including records of proximity and grooming during hour-long observation sessions of adult males (1998-2022) and social networks created from these data.
Files and variables
File: chimp_behav_data.csv
Description: social behavior of mature male chimpanzees sampled at 10-minute intervals
Variables
-
date: date of observation session
-
code: two or three-letter code for chimp
-
prox2: other chimps within 2 meters of the focal subject
-
prox5: other chimps within 2 meters of the focal subject
-
gdyad: grooming dyad
-
party: other mature males observed during 1 hour following the session
-
year: year
-
season: field season
-
grooming: which chimps were involved in grooming
-
groomer: which chimp was giving grooming
-
groomee: which chimp was receiving grooming
-
focal_id: unique code for focal session
-
scan_id: code for 10-minute scans
"NA" represents cases where there are no values for proximity, grooming, or association.
File: learned_graphs_csv.zip
Description: compressed folder containing CSVs for yearly adjacency matrices based on the learned weights, which combined grooming, proximity, and party association (n = 219 total individuals)
File: combined_membership_labels.csv
Description: a file for all 219 individuals that appear in the full network and what group they end up being party of following the permanent fission
- Index: a numeric identifier for each individual chimpanzee
- Name: two- or three-letter code or name for each individual chimpanzee
- Combined Membership Label
File: 77_focal_graph.zip
Description: compressed folder containing CSVs for yearly adjacency matrices based on the male proximity network (n = 77 individuals)
File: 0_threshold.zip
Description: compressed folder containing CSVs for yearly adjacency matrices based on the learned weights, which combined grooming, proximity, and party association (n = 219 total individuals) with a threshold of 0 (indicating whether pairs of chimpanzees had any social tie in a given year).
File: patrol-data-quarterly.csv
Description: counts of boundary patrols each quarter between the Western and Central groups
Variables
- Year
- Quarter: each year is split into four quarters (Q1, Q2, Q3, Q4)
- Group: patrols by the Western group or Central group
- Period: a value combining the year and quarter
- PeriodDate: a snapshot date for each quarter
- Count: total number of patrols
File: population_snapshots.csv
Description: summary of group size across time
Variables
- community: population of the one group before the fission (Ngogo) or of the post-fission groups after the fission (West or Central)
- age_class: infant (0 to 5 years), juvenile (>5 to 9 years), young adolescent (y_adolescent, 9 to <12 years), late adolescent (l_adolescent, 12 to 15 years) or adult (>15 years)
- sex: male or female
- count: number of individuals
- year
- snapshot_date: midpoint date of July 1 used to calculate snapshots of population size each year
Code/software
Files are stored as CSVs.
We used R and Python for data analysis.
Access information
Other publicly accessible locations of the data and related analyses:
- Ren, J., KEI, Y. L., MADRID PADILLA, O. H., & Sandel, A. (2026). Dataset and Code for Network Analysis in "Lethal Conflict After Group Fission in Wild Chimpanzees". In Lethal conflict after group fission in wild chimpanzees. Zenodo. https://doi.org/10.5281/zenodo.18626723
- Sandel, A., Lee, K. C., Angedakin, S., Birungi, C., Kanweri, D., Kalunga, D., Aliganyira, C., Nakayima, P., Kamugyisha, B., Mbabazi, G., Akamumpa, E., Tumusiime, A., He, Y., Reddy, R. B., Negrey, J., Watts, D., Mitani, J., & Langergraber, K. (2026). Space Use Analysis for "Lethal conflict after group fission in wild chimpanzees" [Data set]. In Lethal conflict after group fission in wild chimpanzees. Zenodo. https://doi.org/10.5281/zenodo.18603419
- He, Y., Sandel, A., Wipf, D., Cucuringu, M., Mitani, J., & Reinert, G. (2026). Dataset and Code for Learned Network Analysis in "Lethal Conflict After Group Fission in Wild Chimpanzees". Zenodo. https://doi.org/10.5281/zenodo.19210783
Systematic behavioral data were collected by JCM during 2-3 months during 24 field seasons from 1998-2022. For several analyses, we focus on this male network, as these individuals were well-sampled and adult males are the main participants in territorial behavior and lethal aggression. To create a more thorough picture of the social network, we also included data on males and females that were not systematically followed but interacted with those that were. To construct this full network, we used data on grooming, different distances of proximity, and party association, and implemented a novel optimization approach based on machine learning principles to create a single combined network (developed by YH). We used change-point detection models designed for longitudinal network data (developed and implemented by OMP, YLK, and JR). To determine how individual ranging patterns changed around the time of the major network change, a team of research staff and researchers followed chimpanzees daily throughout the entire year from 2012 to 2022. A large research team of students, staff, and professors with the Ngogo Chimpanzee Project made observations of lethal attacks by chimpanzees from 2018 to 2024.