This paper focuses on the work of Louis Scott and co-workers on the palynology of present-day and past savanna fossil-pollen sites in southern Africa. Evidence for the persistent though fluctuating presence of the pollen of extant species of trees indicates species survival under marked variations in climate over the past 200 ka, but variable assemblage of communities over time, at any one site. This, with contemporary information on the genetics and demographics of savanna tree species, requires that ecologists today must see the savanna as an individualistic, ecogenetic or eco-evolutionary community. Further, present-day savanna communities are the outcome of recent and older history, requiring that ecology must encompass the methods of historical science.