Merging distinct research areas can be synergistic. Arctostaphylos species (Ericaceae) produce fruit that are dispersed by scatter-hoarding rodents. Species also vary greatly in their fruit/seed size. The 2-10 seed per fruit are contained inside nut-like endocarps (nutlets) and remain dormant until stimulated by fire. Consequently, their persistent soil seed banks are the result of decades of interaction between plant fruit production and rodent seed predation with seed that may be decades old. Regardless of variation in seed size, if rodents behave according to optimal foraging theory, this should result in a seed bank density that reflects an energetic giving-up-density (summing the energy in seed remaining after foraging) for the rodent community. If seeds contain the same amount of energy per gram, we predict that seed bank density should be negatively correlated with seed size. We also assume selection will drive plant responses to foraging. Because rodents respond to quality and quantity, we predict that other aspects of fruits and seeds will vary to influence rodent behavior. We suggest that ‘quality’ will reflect the proportion of nutlets with viable seed and that ‘quantity’ will reflect variation in the fusion of multiple endocarps.
Results/Conclusions
To assess whether rodents may be influencing aspects of Arctostaphylos seed banks, soil seed banks from 6 species that vary in seed size were collected and hand sorted from soil. We found a significant negative correlation of seed bank densities with seed size. Fresh fruit were found to contain a large proportion of empty endocarps. This occurred in all species examined. Endocarp fusion also was variable, both within and among fruit for each species. These features of fruit and seed production we interpret as reducing the magnitude of seed predation from caches by increasing handling time and increasing unpredictability in rewards. These data suggest that seed bank dynamics in Arctostaphylos is the result of long-term selection on interactions between an evolving plant lineage and the rodent community. The net result is that selection has modified interactions between rodent communities and plant fruit and nutlet characteristics such that seed banks are predictably sufficient to replace populations killed by fire (obligate seeders) or partially killed by fire (facultative seeders).