Vocal consensus building for collective departures in wild western gorillas
Data files
Sep 10, 2024 version files 1.18 MB
-
data_figure1_GLMM1.csv
33.76 KB
-
data_figure2_GLMM2.csv
12.91 KB
-
data_figure4_GLMM3.csv
10.38 KB
-
data_figure5_S4_S7_Survival1_2.csv
567.20 KB
-
data_Survival1_2_reduced_dataset.csv
361.99 KB
-
elo_functions.R
12.32 KB
-
Figure1_GLMM1.R
10.86 KB
-
Figure2_GLMM2.R
8.21 KB
-
Figure3_data.csv
49.68 KB
-
Figure3_data.R
1.09 KB
-
Figure4_GLMM3.R
7.32 KB
-
Figure5_S4_S7_survival1_2.R
40.94 KB
-
Figure6_data.csv
513 B
-
Figure6.R
4.86 KB
-
FigureS2.csv
30.74 KB
-
FigureS2.Rmd
14.62 KB
-
FigureS3_data.csv
1.83 KB
-
FigureS3.R
729 B
-
FigureS5_data.csv
953 B
-
FigureS5.R
919 B
-
README.md
6.76 KB
Abstract
The ability to coordinate actions is of vital importance for group-living animals, particularly in relation to travel. Groups can only remain cohesive if members possess a cooperative mechanism to overcome differences in individual priorities and social power when coordinating departures. To better understand how hominids achieve spatio-temporally coordinated group movements, we investigated vocally-initiated group departures in three habituated groups of western gorillas (Gorilla gorilla) in the Central African Republic. The large sexual dimorphism of gorillas has led to the untested assumption that the silverback males are the sole decision-makers in gorilla groups although there are also observations that suggest otherwise. To address this, we analysed the direction and timing of group departures and found that high-ranking individuals (silverbacks and high-ranking females) were more successful in indicating the direction of future travel than others, but that the timing of departure was the apparent result of a cumulative vocal voting process amongst all adult group members. Our findings illustrate that even in species with a large sexual size dimorphism, travel decisions can be taken collectively via a consensus-building process.
README: Vocal consensus building in wild western gorillas dataset
https://doi.org/10.5061/dryad.ttdz08m5t
The data were collected on three gorilla groups (N=14 adult individuals) during a 1-year field study in Central African Republic.
The goal of this study was to investigate decision-making and leadership in western gorillas. We focussed on group departures after resting.
To this study belong the following data and codes:
Figure 1
data: data_figure1_GLMM1.csv #contains all-occurrence data with for each subject present at the resting site whether they initiated or not
code: Figure1_GLMM1.R
columns explanations:
- group = study group
- subject = code for individual
- subject_date = subject linked with the date (dd.mm.yyyy)
- nr.id.present = number of individuals (>2y) present at resting site
- id.was.sb = whether subject was a silverback
- norm.elo = normalized elo score of subject
- id.was.focal = whether subject was focal subject during that day or not
- id.initiated = whether subject initiated during the resting event (y/n)
- age.sex = age.sex class subject (adult female, silverback blackback)
- date = date of the observation (dd.mm.yyyy)
- norm.elo = normalized elo score of subject
Figure 2
data: data_figure2_GLMM2 #contains data on initiations and whether they were successful
code: Figure2_GLMM1.R
columns explanations:
- group = study group
- subject = code for subject -> refers to id initiator
- subject_date = initiator linked with the date (dd.mm.yyyy)
- nr.id.present = number of individuals (>2y) present at resting site
- id.was.sb = whether initiator was a silverback
- succesful = whether the initiation was succesful (y/n)
- age.sex = age.sex class initiator (adult female, silverback blackback)
- date = date of the observation (dd.mm.yyyy)
- norm.elo = normalized elo score of initiator
Figure 3
data: Figure3_data.csv #showing timing of vocalizations
code" Figure3.R
columns explanations:
- timing.from.travel = timing of vocalizations in minutes before or after focal subject departed (max 15 minutes before and 5 minutes after)
- caller.is.focal = whether caller was focal
Figure 4
data: data_figure4_GLMM3 #calling rate of focal individuals at resting event
code: Figure4_GLMM3.R
columns explanations:
- group = study group
- subject = code for focal subject
- date = focal subject linked with the date
- nr.id.present = number of individuals (>2y) present at resting site
- subject.is.silverback = whether focal subject was a silverback
- norm.elo = normalized elo score of subject
- age.sex = age.sex class focal subject (adult female, silverback blackback)
- duration_rest_s = lenght of the resting events (in seconds)
- nr.calls = number of grunt-like calls focal subject produced during the resting event
Figure 5
data: data_figure5_S4_S7_Survival1_2 #Effect of vocalizations on departure probability. Resting events were devided into 1 minute intervals. Only the intervals after 15 minutes were used for analysis.
reduced data set: data_Survival1_2_reduced_dataset #same as data before but with resting events excluded that had unknown callers
code: Figure5_S4_S7_survival1_2.R
columns explanations:
Columns:
- subject = code for focal subject
- interval_count = count of the 1 minute interval.
- subject.called = focal subject vocalized in respective 1-minute interval
- others.called = focal subject did not vocalize in respective 1-minute interval
- adults.called.last.five.min = number of adult individuals that vocalized in the current or any of the previous 4 1-minute intervals - i.e. in the past 5 minutes
- group = study group
- subject.is.silverback = whether focal subject was a silverback
- others.stopped.r = whether other individuals than focal stopped resting in the current interval (does not necessarily mean they departed, but they stood up and moved)
- nr.id.present = number of individuals (>2y) present at resting site
- nr.adult.present = number of adult individuals present at resting site
- event.info = whether resting event was seen untill the end or not (censored)
- date = date of the observation (dd.mm.yyyy)
- norm.elo = normalized elo score of subject
- nr.adult.present.5.min.ago = number of adults that was present during the last 5 minutes
- resting_id = identifyier for resting event
Figure 6
data: Figure6_data.csv #Number of vocalizers over cumulative proportion of departures
code: Figure6.R
columns explanations:
- nr.adult.present = number of adult individuals present at the resting site
- nr.adult.callers = number of adults that vocalized in the 5 minutes before departure
- cumulative_proportion = cumulative proportion of departues, per number of adults present
- n.events = number of resting events
- n.events.total = total number of resting events
Supplementary:
Figure S2
data: FigureS2_data.csv #Data frame with agonistic interactions recorded for Elo-calculations
main code: FigureS2.Rmd #script will automatically source the functions in elo_functions.R
additional functions: elo_functions.R
columns explanations:
- group = study group
- date = date of the agonistic interaction (dd.mm.yyyy)
- time = time of the agonistic interaction
- winner = individual that initiated the aggression
- loser = individual that was the target of aggression
- reaction.intensity = outcome of the interaction: 1 or 2) "loser" displaced or does not return ago 3) "loser" returns aggression (tie) 4) "loser" displaces "winner" (script then switches winner loser)
- courtship = whether ago was likely courtship. Did not make difference for elo to exclude these, so let these cases in to remain oblivious to reason agression
Figure S3
data: FigureS3_data.csv #Visualisation of the length of resting events. contains lenght of all complete (seen from start to end) resting events that data was collected on
code: FigureS3.R
columns explanations:
- resting.minutes = duriation resting event in minutes
Figure S5
data: FigureS5_data.csv #cumulative probability of departure over the cumulative proportion of callers
code: FigureS5.R
columns explanations:
- group = study group
- cumulative_callers = relative number of vocalizers (number of vocalizers/number of individuals present)
- cumulative_proportion = cumulative proportion of departues
- n.events = number of resting events
- n.events.total = total number of resting events
all scripts are annotated.
#Figure2_GLMM1.R; Figure3_GLMM1.R; Figure4_GLMM1.R make use of functions provided by Roger Mundry for diagnostic testing. These are publicly available in the supplementary information of https://doi.org/10.1038/s41598-021-01356-6
Methods
Data have been collected on three groups of western gorillas in the Central African Republic during a 12 month field study.