In semi-arid shrubland of the Negev Desert of Israel, harvester ants (Messor ebeninus) may alter abundance, species richness and compositions of herbaceous plant communities by ecosystem engineering of nest patches (affecting conditions and resources) and by granivory as foragers differentially harvest seeds. The mechanism may be alteration of the seed community entering potential seed banks, through variation in sources, movement, and composition of the seed rain. We hypothesized that the positive effects of dense vegetation on harvester ant nests would outweigh reduction by granivory. We tested the effects of granivory at 20 nests by burying one seed trap at three locations; 1 m from the nest, near a foraging trail and in an unforaged area. To test whether nests were seed sources, vegetation was removed prior to seeding on half the nests. Trapped seeds were recorded during seed dispersal in spring and summer.

Results/Conclusions

During primary dispersal in spring, seeds were more abundant near the uncut nests than at the other two locations. Although seed species composition was different on the nests, trap richness was unaffected by location or treatment. In contrast, seed abundance and richness were higher on cut nests during secondary summer dispersal. In spring and summer, foraging trails and unforaged areas were similar, except for higher equitability on foraging trails compared to nests or unforaged areas. Our results show that ants modify seed movement by ecosystem engineering effects on nests and by granivory away from ant nests in different ways. While intact nest vegetation initially had high abundance and richness of locally deposited seeds, it apparently blocked seed import by wind during secondary dispersal. In contrast, granivory had little effect on seed abundance and species richness of the seed community, but decreased dominance structure by preferential removal of the dominant species.