PRIMARY AND SECONDARY SEED DISPERSAL OF BUSH CHINQUAPIN (FAGACEAE) BY SCATTERHOARDING RODENTS

Abstract

Seed dispersal often occurs in two or more discrete stages, but the effects of the latter stages of dispersal on successful recruitment have seldom been measured. We investigated the relative contributions of primary and secondary seed dispersal of Sierra bush chinquapin (Castanopsis sempervirens; Fagaceae) to seedling establishment. Rodents (mostly yellow pine chipmunks, Tamias amoenus) harvested chinquapin nuts rapidly during a three-week period starting in mid-August and scatterhoarded most nuts (81%) in the ground. Most initial (primary) caches contained one nut (range = 1–5 nuts) buried 10–23 mm deep within 46 m of source shrubs. Rodents subsequently retrieved 95% of these nuts and recached at least 33% elsewhere (secondary caches), at lower densities and up to 64 m from source plants. By winter, 41% and 21% of nuts harvested from shrubs remained in scatter caches in 1999 and 2000, respectively. Only 10 caches (4 primary and 6 secondary caches) present in winter 1999 (7.8%), and no caches from winter 2000, produced seedlings the following spring. All seedlings died during the first growing season because of rodent grazing and desiccation. Despite low seedling survival in this study, chinquapin appears to be dependent upon scatterhoarding rodents for seedling recruitment. Secondary dispersal is an important aspect of chinquapin dispersal. Secondary caches contributed more to chinquapin seedling recruitment than primary caches; 2.5% of primary caches and 5.6% of secondary caches produced seedlings. This study suggests that, for propagules that move repeatedly during dispersal, a clear understanding of the dynamics of seed dispersal can be gained only by following seed movements, determining the transition probabilities between different stages of dispersal, and assessing the contribution of each stage of seed dispersal to plant recruitment.