Cone and seed traits of two Juniperus species influence roles of frugivores and scatter-hoarding rodents as seed dispersal agents

Incomplete recovery of seeds from scatterhoards by granivorous rodents: Implications for plant establishment

Granivorous rodents are important components of ecosystems not only because they consume seeds but also because some aid in seed dispersal through seed-caching behaviors. Some rodents bury seeds in shallow pits throughout territories, called scatterhoards, that individuals recover, pilfer, or transfer to other caches. We suspect some single-seed caches in environments represent missed seeds from reclaiming or pilfering caches. We documented the sloppiness of seed removal from scatterhoards of soapweed yucca (Yucca glauca) seeds by Ord's kangaroo rats (Dipodomys ordii). We quantified the frequency and location of seeds remaining. In an experiment with artificial caches of three sizes, kangaroo rats harvested 51% of caches after one night, and 53% had incomplete recovery with at least one seed remaining. The greater the number of seeds in caches, the greater frequency of incomplete recovery. In another experiment with natural and artificial caches, 75% of caches were excavated after 8 days, with at least 70% having at least one seed remaining. Regardless of original cache size, a single seed represented the mode for seeds remaining. Incomplete recovery of seeds likely benefits plant establishment, potentially significantly in some systems. Remaining seeds, especially those buried at bottoms of caches, likely will stay undetected in landscapes, yielding propagules for subsequent plant generations. Soapweed yucca has large but light, flat wind-dispersed seeds, and removal of caches with smaller seeds might have greater frequency of missed seeds during recovery and pilfering by rodents. Our results suggest that scatter-hoarding granivores also contribute to plant establishment by leaving limited numbers of seeds behind when removing caches, at least in some systems.

Seed traits and rodent community interact to determine seed fate: evidence from both enclosure and field experiments

Animal-mediated seed dispersal is an important ecological process in which a strong mutualism between animals and plants can arise. However, few studies have examined how a community of potential seed dispersers interacts with sympatric seed trees. We employed a series of experiments in the Qinling Mountains in both semi-natural enclosure and the field to assess the interactions among 3 sympatric rodent species and 3 Fagaceae tree seeds. Seed traits all showed similar tannin levels but markedly different physical traits and nutritional contents. We found that seeds with heavy weight, thick coat, and high nutritional contents were less likely to be eaten in situ but more often to be eaten after dispersal or hoarded by rodents. These results support both the handling time hypothesis and the high nutrition hypothesis. Surprisingly, we also found that rodents, maybe, preferred to consume seeds with low levels of crude fiber in situ, and to harvest and hoard those with high levels of crude fiber for later consumption. The sympatric rodent species, Cansumys canus, the largest rodent in our study, harvested and hoarded more Quercus variabilis seeds with high physical and nutritional traits, while Apodemus draco, the smallest rodent, harvested more Q. serrata seeds with low physical and nutritional traits, and Niviventer confucianus harvested and hoarded more Q. aliena seeds with medium physical and nutritional traits. Our study demonstrates that different seed traits play different roles in influencing the seed fate and the shaping of mutualism and predation interactions within a community of rodent species.

Pilfering of western juniper seed caches by scatter-hoarding rodents varies by microsite and canopy type

Scatter-hoarding rodents store seeds throughout their home ranges in superficially buried caches which, unlike seeds larder-hoarded in burrows, are difficult to defend. Cached seeds are often pilfered by other scatter-hoarders and either re-cached, eaten or larder-hoarded. Such seed movements can influence seedling recruitment, because only seeds remaining in caches are likely to germinate. Although the importance of scatter-hoarding rodents in the dispersal of western juniper seeds has recently been revealed, the level of pilfering that occurs after initial burial is unknown. Seed traits, soil moisture, and substrate can influence pilfering processes, but less is known about how pilfering varies among caches placed in open versus canopy microsites, or how cache discovery and removal varies among different canopy-types, tree versus shrub. We compared the removal of artificial caches between open and canopy microsites and between tree and shrub canopies at two sites in northeastern California during late spring and fall. We also used trail cameras at one site to monitor artificial cache removal, identify potential pilferers, and illuminate microsite use by scatter-hoarders. Removal of artificial caches was faster in open microsites at both sites during both seasons, and more caches were removed from shrub than tree canopies. California kangaroo rats were the species observed most on cameras, foraging most often in open microsites, which could explain the observed pilfering patterns. This is the first study to document pilfering of western juniper seeds, providing further evidence of the importance of scatter-hoarding rodent foraging behavior in understanding seedling recruitment processes in juniper woodlands.