Seed weight variability of antelope bitterbrush (Purshia tridentata: Rosaceae)

Reproductive output in plants can be affected by a number of factors that can have lasting fitness consequences. Factors influencing seed weight in herbaceous plants have been thoroughly explored, but are relatively unknown in woody plants. This study provides data on variation in seed weight from over 240 antelope bitterbrush plants from 10 sites in the dry interior of British Columbia, Canada. I determined the relationship between seed weight and shrub and site characteristics. Seed weight varied over 9-fold, from 5 to 46 mg, and was skewed in favour of smaller seeds. Most of the variation in seed weight was attributable to variation within individual shrubs (63.2%) whereas different shrubs accounted for 29.1%. Different shrubs produced seeds of different mean weight as did different sites but no measured site or shrub characteristics were related to these seed weight differences. Average seed weight per shrub was not associated with shrub size or percent dieback. Average seed weight per site was not associated with browse intensity, density of knapweed or cheatgrass, shrub density, percent cover of bare soil, or cation exchange capacity of the soil. Seed weights of some antelope bitterbrush plants varied more than others; coefficient of variation ranged from 9 to 39%. Seed weight variability was positively correlated with the number of twigs browsed during the winter previous to seed production. Antelope bitterbrush is known to overcompensate twig production in response to browsing, and to forfeit the production of flowers in favour of twigs. I discuss how this might result in the development of more varied seed weights.