Effect of weed patch size on seed removal by harvester ants

Abstract

In dryland cereals in North-eastern Spain, the harvester ant, Messor barbarus L., is responsible for removal of a large proportion of the newly produced weed seeds (40-100%). The probability that seeds will be found by the ants may be influenced by weed patch size. To investigate this source of variability, 30 seed patches were created in each of three, 50 × 50 m, blocks in a cereal field after harvest, by sequentially seeding (10, 16 and 17 August 2010) with 2000 seeds m-2 of Avena sativa L.. Patch size varied from 0.25 to 9 m2. Twenty four hours after seeding, the remaining seeds were collected and seed removal rates estimated. Average seed removal rate was lowest in the smallest (78-94%) and highest in the largest patches (99-100%). Differences were mainly caused by the fact that some of the smaller patches (9.7%) were not found. However, when patches were found, they were exploited at equal rates (98-100%). As predicted, the probability of finding a patch increased slightly, but significantly, with increasing patch size. When a patch was found, it was almost always fully exploited, resulting in very high seed removal rates, irrespective of patch size. These results indicate that the size of the seed patch is only a minor source of variation influencing this form of biological control of weeds.

Keywords

• biological control
• encounter rate
• granivory
• harvest ants
• patch size
• seed predation