Did the elephant and giraffe mediate change in the prevalence of palatable species in an East African Acacia woodland?

We report on a 2009 survey of Acacia woodlands in the Seronera area of central Serengeti, Tanzania, and compare the results to previous surveys from the 1970s-1980s. We document a substantial change in woodland structure and composition. From 1978 to 2009, woody plant density (mean ± SD) declined from 255 ± 35 trees ha^-1 to 147 ± 14 trees ha^-1. Canopy cover declined from 15.1 ± 1.9% to 5.7 ± 0.6%. Canopy volume declined from 1810 ± 207 m³ ha^-1 to 1410 ± 121 m³ ha^-1. A 19% increase in the relative canopy volume of Acacia robusta, a species avoided by the giraffe, represents a sharp decline in the quality of the giraffe's food supply, which was last comprehensively assessed in 1978. We examine these changes in the context of Pellew's 1983 woodland dynamics model describing the impacts of elephant, giraffe and fire on an Acacia tortilis tree population and determine that our data are qualitatively consistent with this model. We hypothesize that selective elephant and giraffe browsing has contributed to an increase in the relative dominance of unpalatable species, consistent with Pellew's predictions.