The roles of seed eating ants and rodents in North American deserts have been well established. These granivores are the primary consumers of seeds of annual and perennial plants, and through their distinctive foraging behaviors affect seed distribution and density. However, the interactions between granivores and seeds in habitats disturbed by wild fire and their roles in post-fire vegetation recovery are poorly understood. To elucidate seed predation patterns and potential impacts on the seed bank by ants and rodents, we carried out a seed removal experiment in burned and unburned creosote (Larrea tridentata) scrub communities in the Mojave Desert in southern Nevada, USA. We studied seed removal patterns using wire cages that either excluded or included ant or rodent access to seeds. Granivory treatments were completely randomized within six experimental blocks located in burned and unburned habitats.  We offered 20 seeds of eight native species, and quantified seed removal by subtracting the number of seeds that remained in a Petri dish after a four day trial. We used generalized linear models to analyze the effects of granivory treatments and fire disturbance on seed removal.

Results/Conclusions

Our results indicate that seed removal was higher in unburned (mean ± SE, 29.28 ± 12.60 seeds removed) than in burned (21.61 ± 5.04 seeds removed) habitats. Mean number of seeds removed by rodents in unburned (62.83 ± 35.70 seeds) was significantly greater from burned (15.17 ± 6.21 seeds) habitats. Ants removed a higher number of seeds in burned (41.33 ± 9.88 seeds) than in unburned (13.67 ± 2.51 seeds) habitats; however, this pattern was not significant. It appears that rodents are removing more seeds of a large-seeded species, Coleogyne ramosissima, (seed mass, 15.86 ± 0.61 mg) while ants are selecting seeds of small-seeded species, Penstemon bicolor [0.54 ± 0.11 mg], Encelia farinosa [1.00 ± 0.10 mg], and Sphaeralcea ambigua [1.09 ± 0.08 mg]. Our results suggest that granivorous ants and rodents are not playing a significant role in structuring plant communities in burned areas.  Ant and rodent densities may be low due to changes in vegetation structure caused by fire, but these densities will likely recover with time; consequently, seed removal patterns and the activity of granivores are expected to change.