Submitted to: Western North American Naturalist
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/8/2018
Publication Date: 8/9/2018
Citation: Longland, W.S., Dimitri, L.A. 2018. Interaction between seed detectability and seed preference affects harvest rates of granivorous rodents. Western North American Naturalist. 78(2):195-203. https://doi.org/10.3398/064.078.0210.
Interpretive Summary: Production of new seedlings of Indian ricegrass, a grass native to the Great Basin and an important source of forage for the range livestock industry, comes largely from Indian ricegrass seeds that have been cached by desert rodents. Although Indian ricegrass seeds are a highly preferred food for desert rodents, these animals are better able to locate buried seeds of an alternate food plant – cheatgrass. Indian ricegrass has adapted to producing seeds which mask odors that cue rodents to locations of buried seeds. This makes it less likely that Indian ricegrass seed caches will be relocated and consumed by rodents before they have a chance to produce new seedlings. Because cheatgrass is an introduced species, it has not evolved together with North American desert rodents, and it therefore lacks such a seed-masking adaptation.
Technical Abstract: Granivorous rodents commonly exhibit preferences for seeds of particular plant species. However, among buried seeds that are available, rodents may find alternate, less desirable seeds to be more easily located by olfaction than preferred seeds. Seeds of Indian ricegrass (Achnatherum hymenoides) are a highly preferred food resource for seed-caching desert heteromyid rodents. We tested relative abilities of heteromyids to locate buried caches of Indian ricegrass seeds versus seeds of another plant species (cheatgrass, Bromus tectorum) that are common in heteromyids diets at a western Nevada field site. Rodents located more cheatgrass than Indian ricegrass seed caches in the field when seeds were unaltered. When ground, however, volatile compounds released from inside seeds increased the detectability of preferred Indian ricegrass seeds, and more ground Indian ricegrass seed caches were found and removed compared with ground cheatgrass caches. Grinding cheatgrass seeds did not affect harvest rates of cheatgrass caches. Results of a laboratory cache removal experiment using two heteromyid species, Merriam’s kangaroo rats (Dipodomys merriami) and long-tailed pocket mice (Chaetodipus formosus), mirrored field cache removal results. Dispersal of Indian ricegrass seeds and establishment of new seedlings occurs largely through emergence of seedlings from heteromyid scatterhoard caches. Indian ricegrass seeds have been selected for reduced olfactory detectability to minimize the probability that they are recovered by rodents for later consumption once they have been cached.