Seed predation by rodents can strongly influence plant recruitment and establishment. The extent to which predation risk indirectly alters plant survival in tropical forests via impacts on granivory is unclear, making it difficult to assess the cascading impacts of widespread predator loss on tree recruitment and species composition. Experimental field studies that manipulate predation risk can help address these knowledge gaps and reveal whether antipredator responses among small mammals influence plant survival. We used camera traps and seed predation experiments to test the effects of perceived predation risk (via predator urine gel) on foraging behaviour of and seed removal by murid rodents in an unlogged and unhunted rainforest in Malaysian Borneo. We also explored the influence of seed traits (e.g., seed size) on removal by granivores and assessed whether granivore preferences for particular species were affected by predator urine. Murid visits to seed plots were positively related to overall seed removal, but were not affected by predator scent. Granivory was the lowest for the largest-seeded (>6 g) plant in our study, but was not influenced by predation risk. Predator urine significantly affected removal of one seed taxon (Dimoocarpus, ∼0.8 g), suggesting that removal by granivores may be affected by predation risk for some seed species but not others. This could have implications for plant species composition but may not affect the overall level of granivory.

Although seed-size preference by granivores is relative to the sizes of seeds offered by researchers, preferences are often inappropriately discussed and interpreted as if they are for absolute seed sizes. Granivores often prefer relatively small seeds in forests and relatively large seeds in herbaceous communities, although the absolute size of preferred seeds in each is often similar. These results are consistent with predictions of optimal foraging theory. Researchers are encouraged to take heed of optimal foraging theory and both relative and absolute seed size when interpreting seed-size preferences.

The influence of farm management practices on ground-dwelling natural enemy communities and predation of insects and weed seeds was investigated over the first 2 years of the transition from conventional to organic production. Three transition strategies were selected that differed in their management and input intensities, and were characteristic of pasture/ley systems (low intensity), cash grain systems (intermediate intensity), and vegetable production (high intensity). Beneficial arthropods (insectivores and granivores) were monitored using pitfall (arthropod activity) and quadrat (arthropod density) samples. The frequency of predation on restrained larvae of Galleria mellonella and the species observed feeding were recorded. Weekly removal rates of weed seeds representative of abundant species at our site were monitored over a 3-week period during fall. Management intensity affected the activity and abundance of biological control agents. In year two of the transition, biological control agent densities were higher in the low-intensity treatment than in the other two treatments, but activity of insectivores and granivores was reduced in this treatment relative to the higher intensity systems. The patterns in the abundances of biological control agents may be explained by habitat stability within the different cropping systems. Quadrat samples were strongly correlated with the insectivory index, although pitfall samples were not. Insectivory rates were highest (>80% of G. mellonella larvae) in the low-intensity treatment. Predation patterns over a 17-h period differed substantially among the management treatments, indicating behaviorally distinct insectivore communities. Seed removal was also highest in the low-intensity treatment. We conclude that low-intensity cropping systems are most favorable to the abundance and function of beneficial ground-dwelling arthropod communities (insectivores and granivores) during the transition process.