Methods
Individual seed production data from 59 populations of 24 species of plants was obtained. All available datasets that reported individual plant data were used and, within these studies, we included all individuals that reproduced at least once during the course of the study. Studies ranged from 4 to 22 years in length. Methods for estimating seed productivity ranged across studies. In order to investigate the potential for selection acting on different components of masting, we compared plant species according to their growth form (herbaceous or woody) and their primary pollen vector (animal or wind/gravity). We were unable to effectively compare species according to their primary seed dispersal vector because of poor representation in our data of species dispersed by mutualistic frugivores.
More detailed information regarding the methods and how the data was used appear in the publication, 'Dissecting components of population-level variation in seed production and the evolution of masting behavior', Walter D. Koenig et al., Oikos 102:581-591, 2003.