Seed protection through dispersal by African savannah elephants (loxodonta africana) in Northern Tanzania

Abstract

Seed dispersal by animals is important for maintaining healthy populations of many tree species. The Janzen-Connell hypothesis, states that trees are under selective pressures to the have their seeds dispersed away from the parent plant and into an environment more suitable for growth. Seeds typically do not survive underneath the parent plant for a myriad of reasons, including light and nutrient limitations, and excessive predation. Large-seeded tree species are especially affected by these factors because their seeds cannot be dispersed by abiotic factors, such as wind. Trees with large seeds that can only be effectively dispersed by largebodied animals are referred to as megafaunal syndrome species. African forest elephant disperser effectiveness has been well studied. African savannah elephants may fill a similar niche, although experimental data are few. African savannah elephants have been suggested as critical seed dispersers, and may be the only remaining organisms capable of effectively dispersing seeds of megafaunal syndrome species. I examined the effectiveness of savannah elephant dung as a protective barrier for three tree species: Acacia tortilis, Tamarindus indica, and Balanites aegyptiaca. Experimental treatments were established to measure the effect of dung in protecting passed seeds. I also addressed the Janzen-Connell model. I predicted that seeds in dung and seeds away from the parent tree would experience less infestation than fresh seeds. Simple linear regression was used to determine daily removal and infestation rates. Two-way analysis of variance was used to compare time in days and treatment and their interaction. Multiple comparisons using a Tukey's test of honest significant differences were made to check for true differences between paired treatments from the ANOVA. Finally, loglinear analysis was used to test for differences among infestation of seeds at different distances from adult conspecific trees. In January 2013, ants or termites in Experiment 1 likely removed passed seeds in dung. Fresh seeds experienced similar levels of removal. Chi-square analysis of data supported my hypothesis and revealed differences in beetle emergence between passed and fresh seeds. In support of my prediction, in October 2013, seeds in dung experienced less beetle infestation than fresh seeds in Experiment 2. Seeds at distances greater than five meters experienced less beetle infestation than seeds underneath conspecific trees in Experiment 3. This study was the first to experimentally address post-dispersal seed fate for megafaunal syndrome species by savannah elephants. It creates a link between disperser effectiveness of forest and savannah elephants, and provides foundation for further examining savannah elephants as seed dispersers of megafaunal syndrome species.