Do food distribution and competitor density affect agonistic behaviour within and between clans in a high fission-fusion species?
Data files
Jan 10, 2024 version files 170.31 KB
-
clan-level_BGC_encounters.csv
-
corr_duration_clan-level_BGC.csv
-
correlates_WGCagonism.csv
-
diff_agonism_wcbc_5clans.csv
-
diff_wcbc_allclans_grsize.csv
-
grass_forest_vs_grassland.xlsx
-
grass_kabini.csv
-
NI_I_diff_agonism_wcbc_5clans.csv
-
README.md
Abstract
Socioecological theory attributes social variation in female-bonded species to differences in within- and between-group competition, shaped by food distribution. Strong between-group contests are expected over large, monopolisable resources, but not when low-quality food is distributed across large, undefended home ranges. Within-group contests are expected to be more frequent with increasing heterogeneity in feeding sites. We tested these predictions in female Asian elephants, which show traits associated with infrequent contests – predominant graminivory, overlapping home ranges, and high fission-fusion. We examined how agonistic interactions within and between female elephant clans (social groupings) vary with food distribution and competitor density. We found stronger between-clan contests than that known from neighbouring forests and more frequent agonism between females between clans than within clans. Such strong between-clan contest is attributable to food patchiness as the Kabini grassland in the study area had three times the grass biomass as adjacent forests. Within-clan agonism was also frequent but was not influenced by food distribution, contradicting socioecological predictions. Contrary to recent claims, increasing within-clan agonism with group (party) size showed that ecological constraints operate despite high fission-fusion in Asian elephants. Thus, despite graminivory and fission-fusion, within-clan and between-clan agonism can be frequent, especially at high population density.
README: Do food distribution and competitor density affect agonistic behaviour within and between clans in a high fission-fusion species?
https://doi.org/10.5061/dryad.15dv41p3k
There are eight data files here, including data on a) grass biomass in the Kabini grassland area and Nagarahole forest, b) rates of agonistic interactions between and within females in Asian elephant clans, and c) the ecological correlates of within-clan and between-clan agonistic interactions.
Description of the data and file structure
Description of data files linked with the article:
Gautam H and Vidya TNC. 2023 Do food distribution and competitor density affect agonistic behaviour within and between clans in a high fission–fusion species? R. Soc. Open Sci. 10: 230990
Description of data files linked with the article:
Gautam H and Vidya TNC. 2023 Do food distribution and competitor density affect agonistic behaviour within and between clans in a high fission–fusion species? R. Soc. Open Sci. 10: 230990
1) grass_forest_vs_grassland.xls: Datasets to check habitat-level patchiness i.e. comparison of grass biomass (g/m.sq.; freshly harvested grass biomass, i.e. not oven dried) across habitat types (Kabini grassland vs forests of Nagarahole and Bandipur). There are six sets of grass biomass data from this region: 1) Kabini grassland 2015, 2) Kabini grassland 2016, 3) Nagarahole forests (end of wet season in 2013), 4) Bandipur forests (dry season in 1993), 5) Nagarahole forests (dry season in 2022) and 6) Kabini grassland (dry season in 2022). Summary statistics are provided below each dataset.
In datasets 1, 2 and 3: the monthly average column has values in the last row for each month while other rows are not applicable and hence have n/a.
In datasets 5 and 6: Avg_biomass_plot column has biomass averaged from 4 quadrats in the last row for each plot; other three rows are not applicable and hence have n/a.
2) Code_grass.R: R script to analyze variation in grass abundance (percentage cover, biomass, and height) within the Kabini grassland across different focal zones, months and year (tables in Supplement 1 for the paper). Data file to be used: grass_kabini.csv
3) grass_kabini.csv: data file to analyze variation in grass abundance (percentage cover, biomass, and height) across different focal zones, months and year (tables in Supplement 1 for the paper).
Description of columns in the csv file:
- year = year of measurement (2015 or 2016);
- month = names given to the four 30-day periods (broadly corresponding to Feb, Mar, Apr and May);
- zone = name of the focal zones or large stretches of grasslands (six focal zones);
- plot = ID of the four plot-clusters within each focal zone (see Figure 2c in paper);
- gr_cvr_1m = average percentage cover of grass in the plot-cluster (across 5 constituent quadrats);
- gr_biomass_1m = average grass biomass (g/m.sq.) in the plot-cluster (across 5 constituent quadrats);
- gr_avheight_1m = average grass height (cm) in the plot-cluster (across 5 constituent quadrats);
- StdDev - cover = standard deviation in grass cover in the plot-cluster (across 5 constituent quadrats);
- StdDev - biomass (g/m.sq.) = standard deviation in grass biomass (g) in the plot-cluster (across 5 constituent quadrats);
- StdDev - Mean height (cm) = standard deviation in grass height (cm) in the plot-cluster (across 5 constituent quadrats);
- CV - cover = coefficient of variation (%) in grass cover in the plot-cluster (across 5 constituent quadrats);
- CV - biomass (g/m.sq.) = coefficient of variation (%) in grass biomass in the plot-cluster (across 5 constituent quadrats);
- CV - Mean height (cm) = coefficient of variation (%) in grass height in the plot-cluster (across 5 constituent quadrats).
In row 73: the biomass values for plot D show the imputed means calculated from three other plots (A, B and C) in the zone since the abundance data for plot D could not be collected. The subsequent columns have n/a because the variability indices could not be calculated due to the lack of data from 5 constituent quadrats in plot D.
4) Code_WGBG_diff_allclans.R: R script to 1) compare the rates of within-clan and between-clan agonism, keeping clan-identity as a random effect and 2) compare the slope of the relation between competitor density and rate of agonism for within- and between-clan agonism (this analysis was done using simple regression for each type and then comparing the slopes, see formula in Supplement 3 of this paper). Data file to use: diff_wcbc_allclans_grsize.csv
[For Fig. S11 and Table S6 in the Supplement, there is a similar code called
5) “Code_WGBG_diff_5clans_agonism.R” for comparing within-clan and between-clan agonism for the top five common clans. For this script, use
6) diff_agonism_wcbc_5clans.csv for agonism per female (description similar to diff_wcbc_allclans_grsize.csv listed below), and
7) NI_I_diff_agonism_wcbc_5clans.csv This file is to analyze NI/R ratio (similar description as the diff_wcbc_allclans_grsize.csv file below; excluded rows with n/a in NI_I_ratio column.]
8) diff_wcbc_allclans_grsize.csv: Data file to compare the rates of within-clan and between-clan agonism. Use R script: Code_WGBG_diff_allclans.R
Description of columns in the csv file:
- type_wcbc = type of focal group observation of agonism (within- or between-clan agonism);
- year = Year of observation;
- month = Names given to the four 30-day periods that roughly correspond to February, March, April and May;
- plot = ID of the unique plot-cluster near which agonims occurred;
- focal_sno = ID of the focal group observation;
- AF_sum_grsize = local competitor density (adult female group size for within-clan agonism, and sum of adult female group size of the two competing groups for between-clan agonism);
- date = date of observation;
- duration_obs = duration of observation (minutes);
- I_count = total no. of independent individual-level agonistic interactions (see Figure 1 in the paper);
- agonism_per_h = rate of agonism experienced per hour by an average female in the group (see Supplement table S2 and fig S7 for calculation);
- NI_count = total no. of non-independent agonistic interactions (i.e. the interactions occurring within 15 min of other interactions) observed in the focal;
- NI_I_ratio = NI:I ratio (number of non-independent interactions per independent interaction);
- clan_id = ID of the clan observed;
- clan_id_bg = ID of the clan-combination for between-clan agonism (valid only for between-clan agonism).
n/a cells in the NI_I_ratio column: when there is no independent interaction (i.e. I_count is 0), NI/I ratio is not applicable.
9) Quadr_Fig_grsize_WGC.R: R script to make Figure 4b
10) Doublescatter_grsize_vs_WGBG.R: R script to make Figure 4a
11) Code_correlates_WGCagonism.R: R script to analyze 1) the correlates of within-clan agonism (Table 1 in the paper), and 2) the quadratic relationship between group size and within-clan agonism.
Data file to be used: correlates_WGCagonism.csv.
12) correlates_WGCagonism.csv: data file used in the code to analyse correlates of within-clan agonism (Table 1 in the paper). R script to be used: Code_correlates_WGCagonism.R
Description of columns in the csv file:
- year = year of observation;
- month = names given to the four 30-day periods that roughly correspond to February, March, April and May;
- zone = name of the focal zones or large stretches of grasslands (we sampled six zones, full names in the supplement);
- Nearest Plot-Cluster = unique ID of the nearest plot cluster (of grass sampling) where the observation happened (~within 100m);
- Focal S.No. = serial number of focal observation;
- no._AF = number of adult females in the group;
- date = date of observation;
- focal_duration_obs = duration of the focal group observation (minutes);
- no._I = total no. of independent agonistic interactions observed in the focal;
- agonism_perAF = rate of agonism experienced per hour by an average female in the group (see Supplement table S2 and fig S7 for calculation);
- no._NI = total no. of non-independent agonistic interactions (i.e. the interactions occurring within 15 min of other interactions) observed in the focal;
- NI_I_ratio = NI:I ratio (not applicable entered in rows with zero I interactions);
- clan_id = ID of the clan observed;
- cover_plot = average percentage cover of grass in the plot-cluster (across the 5 constituent quadrats);
- biomass_plot = average grass biomass (g/m.sq.) in the plot-cluster (across the 5 constituent quadrats);
- height_plot = average grass height (cm) in the plot-cluster (across the 5 constituent quadrats);
- sd_cover_plot = standard deviation in grass cover in the plot-cluster (across 5 constituent quadrats);
- sd_biomass_plot = standard deviation in grass biomass (g) in the plot-cluster (across 5 constituent quadrats);
- sd_height_plot = standard deviation in grass height (cm) in the plot-cluster (across 5 constituent quadrats);
- CV_cover_plot = coefficient of variation (%) in grass cover in the plot-cluster (across 5 constituent quadrats);
- CV_biomass_plot = coefficient of variation (%) in grass biomass in the plot-cluster (across 5 constituent quadrats);
- CV_height_plot = coefficient of variation (%) in grass height in the plot-cluster (across 5 constituent quadrats).
13) Code_corr_glm_clan-levelBGC.R: R script to analyze the rate of clan-level between-clan agonistic encounters in focal zones using a generalized linear model (Poisson structure). Data file to be used: clan-level_BGC_encounters.csv
14) clan-level_BGC_encounters.csv: Data file to analyze the rate of clan-level between-clan agonistic encounters in focal zones. R script to be used: Code_corr_glm_clan-levelBGC.R.
Description of columns:
- Sno. = Serial number;
- date = Date of observation;
- year = Year of observation;
- zone = Name of the focal zones or large stretches of grasslands (we sampled six focal zones);
- month = names given to the four 30-day periods that roughly correspond to February, March, April and May;
- interval_id = time periods for the 2.5-hour intervals during which observations were made;
- area_zone = area of each zone in sq.km.;
- clan_count = Total number of clans present in the focal zone;
- no._AF = Total number of adult females in the focal zone;
- no._BG_encounter = Number of clan-level agonistic between-clan encounters observed in the focal zone;
- zone_cover = average percentage cover of grass in the focal zone (average from the averages of four plot-clusters);
- zone_biomass = average grass biomass (g/m2) in the focal zone (average calculated from the four constituent plot-clusters);
- zone_height = average grass height (cm) in the focal zone (average calculated from the four constituent plot-clusters);
- zone_sd_cover = standard deviation of the percentage cover of grass (across four plot-clusters) in the focal zone;
- zone_sd_biomass = standard deviation of the grass biomass (across four plot-clusters) in the focal zone;
- zone_sd_height = standard deviation of the height of grass (across four plot-clusters) in the focal zone;
- zone_CV_cover = coefficient of variation (%) of the percentage cover of grass (across four plot-clusters) in the focal zone;
- zone_CV_biomass = coefficient of variation (%) of grass biomass (across four plot-clusters) in the focal zone;
- zone_CV_height = coefficient of variation (%) of grass height (across four plot-clusters) in the focal zone.
15) Code_correlates_BGC_duratn.R: R script to analyze the duration of clan-level between-clan encounters. Data file to be used: corr_duration_clan-level_BGC.csv
16) corr_duration_clan-level_BGC.csv: Data file to analyze the duration of clan-level between-clan encounters. R script to be used: Code_correlates_BGC_duratn.R.
Description of columns:
- year = Year of observation;
- month = names given to the four 30-day periods that roughly correspond to February, March, April and May;
- zone = name of the focal zones or large stretches of grasslands (we sampled six focal zones: the full names are in the supplementary material associated with this paper;
- focal_s_no = ID of the focal group observation;
- sum_AF = sum of adult female group size of the two competing groups;
- diff_AF = difference in the adult female group size of the two competing groups;
- date = date of observation;
- BGD_duration_min = duration of the between-clan encounter (minutes);
- no._I = total no. of independent individual-level between-clan agonistic interactions observed in the focal (see Figure 1 in the paper);
- no._NI = total no. of non-independent individual-level between-clan agonistic interactions (occuring within 15 min of another interaction) observed in the focal;
- Clan_combinat = ID of the combination of the two competing clans;
- plot_id = unique ID of the nearest plot cluster (of grass sampling) where the observation happened (~within 100m);
- plot_cluster = ID of the plot-cluster within focal zone;
- cover_plot = average percentage cover of grass in the plot-cluster (across the 5 constituent quadrats);
- biomass_plot = average grass biomass (g/m.sq.) in the plot-cluster (across the 5 constituent quadrats);
- height_plot = average grass height (cm) in the plot-cluster (across the 5 constituent quadrats);
- sd_cover_plot = standard deviation in grass cover in the plot-cluster (across 5 constituent quadrats);
- sd_biomass_plot = standard deviation in grass biomass (g) in the plot-cluster (across 5 constituent quadrats);
- sd_height_plot = standard deviation in grass height (cm) in the plot-cluster (across 5 constituent quadrats);
- CV_cover_plot = coefficient of variation (%) in grass cover in the plot-cluster (across 5 constituent quadrats);
- CV_biomass_plot = coefficient of variation (%) in grass biomass in the plot-cluster (across 5 constituent quadrats);
- CV_height_plot = coefficient of variation (%) in grass height in the plot-cluster (across 5 constituent quadrats);
- cover_zone = average percentage cover of grass in the focal zone (average from the averages of four plot-clusters);
- biomass_zone = average grass biomass in the focal zone (average calculated from the four constituent plot-clusters);
- height_zone = average grass height in the focal zone (average calculated from the four constituent plot-clusters);
- sd_cover_zone = standard deviation of the percentage cover of grass (across four plot-clusters) in the focal zone;
- sd_biomass_zone = standard deviation of the grass biomass (across four plot-clusters) in the focal zone;
- sd_height_zone = standard deviation of the height of grass (across four plot-clusters) in the focal zone;
- CV_cover_zone = coefficient of variation (%) of the percentage cover of grass (across four plot-clusters) in the focal zone;
- CV_biomass_zone = coefficient of variation (%) of grass biomass (across four plot-clusters) in the focal zone;
- CV_height_zone = coefficient of variation (%) of grass height (across four plot-clusters) in the focal zone.
Sharing/Access information
Links to other publicly accessible locations of the data: None.
Data was derived from the following sources: None.
These are primary data collected from the field.
Code/Software
The R codes used to analyse data are also attached. Descriptions are above.