Fine-scale genetic structure and inferred sociality in African forest elephants

Background/Question/Methods

Some species live in flexible groups to maximize the fitness benefits of group living, while minimizing costs through changing group size and structure. African savanna elephants (Loxondonta africana) have a sophisticated fission-fusion social structure composed of hierarchical tiers of matrilines and influenced by resources.  African forest elephants (Loxodonta cyclotis) are ecologically and genetically distinct, yet it is unknown how ecological differences influence social structure.  We examined fine-scale genetic structure of adult female forest elephants to elucidate patterns of sociality. We collected dung samples for genetic analyses during four one-week sessions in Lopé National Park, Gabon in 2008 and 2010. GPS coordinates were taken upon dung collection and bolus size was measured to determine an age class for the individual.  Samples were considered to be in the same group if they found on the same day, were of the same freshness, and within 100 m of each other.  Samples were genotyped at 10 microsatellite loci and genetically sexed.  We used spatial autocorrelation analyses and Mantel tests to investigate the relationship between relatedness and spatial proximity, and permutation tests and social networks to investigate group structure.  We analyzed our data at three temporal scales: one day, one week, and within years. 

Results/Conclusions

Spatial autocorrelation analyses revealed significant positive genetic structure in two of the four sampling sessions, the single day analysis, and both years within the first distance class (0-5.0 km), meaning that individuals were more closely related to each other than expected by chance within these distances.  One session, the single day analysis, and year 2008 had a significant negative relationship in the Mantel tests.  As geographic distance increased, overall relatedness decreased in all sessions, although not all results were significant.  Individuals within groups were significantly more related to each other than to individuals between groups.  The results of the social networks suggest more extensive matrilineal groups than detected from dung sampling and previous observational studies, and most groups were of the same matriline.  Our results do not violate the assumptions of a fission-fusion model for African forest elephants.