Bulot, A., Provost, E. & Dutoit, T.

Refuse pile turnover by harvester ants (Hymenoptera: Formicidae) increases seed density and seedling species richness in dry grasslands

Messor barbarus (Linnaeus, 1767) ants are Mediterranean region seed predators. However, transported seeds can be found rejected in refuse piles around their nests, making this ant a seed dispersal agent. This raises the following questions: Do refuse piles affect seed distribution? Do their small-scale seed composition, richness, and density in autumn and winter differ from those of areas without refuse piles? Are there differences in the vegetation found the following spring in refuse piles that have survived the winter or on sites where they have been destroyed, compared with control areas without refuse piles? In a Mediterranean steppe, we measured autumn and winter seed banks in a greenhouse, both in refuse piles and in controls. The following spring, in situ seedlings from refuse piles, from refuse piles artificially destroyed (sieved) in winter, and from controls without refuse piles were recorded. Seedling species richness and density were significantly higher in autumn in refuse piles than in controls. Nevertheless, no increased seedling contribution from the transient seed bank was detected in the winter refuse piles. The following spring, natural persistent refuse piles showed no seedlings. However, seedling species richness and density were significantly higher in places where refuse piles had been sieved before in winter. The construction of refuse piles by M. barbarus locally concentrates seed density and seed species richness; for the first time, a positive impact on seedlings is observed in places where refuse piles were sieved before winter and the potential bias of this methodology must be now compared with actions of natural agents (surface runoff) that destroyed refuse piles in winter.