The socioecology of elephants: Analysis of the processes creating multitiered social structures
Data files
Oct 16, 2024 version files 369.70 KB
-
dist.matrix.01-04.csv
141.16 KB
-
dist.matrix.98-01.csv
87.59 KB
-
dist.matrix.Jun12.Jul14.csv
137.40 KB
-
README.md
3.56 KB
Abstract
In this paper, we investigate the formation and function of the multilevelled, fission–fusion social structure in a free-ranging African elephant, Loxodonta africana, population. We quantitatively identified the existence of four social tiers by using cluster analysis on individual association data. We assessed the effects of season and study period on social structuring and levels of cohesion within and among social units. We found that second-tier units, potentially the equivalent of the ‘family’, were stable across seasonal periods but the number of units increased as the study progressed and the population grew. It appears that these units were sufficiently small not to be influenced by ecologically related factors, such as resource competition, that might otherwise lead to them splitting. On the other hand, third- and fourth tier units were significantly affected by season in a way that suggests a trade-off between ecological costs (e.g. from resource competition) and different social and ecological benefits (e.g. from predator defence, territoriality, knowledge sharing and rearing of young). Age structure also appeared to influence this multitiered social organization. The size of second-tier social units was significantly affected by the age of matriarchs: units lead by matriarchs likely to be grandmothers (i.e. females 35 years and older) were significantly larger than those lead by younger matriarchs. We present a conceptual framework for understanding the emergence of multiple-tier social structure from interactions driven by socioecological processes. This study is the first to use rigorous quantitative methods to statistically show the existence of four hierarchical tiers of social organization in a nonhuman animal. Additionally, our results elucidate the role that ecological processes play in producing complex social structures.
https://doi.org/10.5061/dryad.djh9w0w6f
Description of the data and file structure
Three files containing distance matrices (1 - simple association indices) of adult female elephants. The three matrices represent different time periods: June 1998 - May 2001, June 2001 - August 2004, June 2012 - July 2014. These matrices were analyzed using clustering algorithms to derive social networks and social trees capturing relationships between individual elephants. These data can be used to assess social network structure and node level social network metrics of wild African elephants.
Description of the data
The data provided here are association distance matrices (i.e. 1 - simple ratio association index values) for successive, multi-year time periods. Data are presented in matrices where individual identity codes are listed across column 1 elements and row 1 elements in the matrices. Individual identify codes are unique letter number combinations assigned to every identified elephant in the population (e.g., R1 - S92.94). NA values represent elements for which no data exit because the two individuals were not available to interact during the time period due to the death of one individual before the recruitment of the other. Association distance values were derived from observations collected as described below.
The majority of observations were collected by observers recording all individuals encountered along one of four set routes travelled daily within the study area. Routes followed existing roads running the length of the Samburu and Buffalo Spring National Reserves in northern Kenya, both to the north and south of the Ewaso N'giro River. Routes 1 and 2 were within 500 m of either riverbank and routes 3 and 4 were located approximately 5 km from the river on either side. Observations of elephant aggregations during nontransect movements in the study area (opportunistic sampling) were also recorded and used in this study. All observations were recorded following the same protocol. Once elephants were spotted, observers approached to within 100 m or less of the aggregation and recorded the following data: (1) identity of individuals present; (2) accuracy of the observation (recorded in categories: (a) 100% identification of all individuals; (b) identification of all breeding females and males excluding calves; and (c) incomplete identification); (3) location of aggregation; and (4) the date, time, observer name and route name. The observer stayed with the aggregation until all individuals were recognized, unless thick vegetation did not allow a complete accounting.
Description of the file strucutre
Each file represents the distance matrices derived from data collected over the successive years in the matrix name:
June 1998 - May 2001
June 2001 - August 2004
June 2012 - June 2014
Association distance matrices are organized where the name of each individual elephant inhabits the elements of Row 1 and Column 1, and successive elements in the matrix represent the association distance of the corresponding row 1 and column 1 constituents. NA values represent pairs that did not overlap during the data sampled time period (e.g., individual 1 died before individual 2 was recruited into the sample).
Sharing/Access information
Please contact George Wittemyer g.wittemyer@colostate.edu for additional information.