Questions: In a system of five annual plant species restricted to nest-mounds of the ant Lasius flavus in a perennial grassland: 1. Are the population dynamics influenced by ant disturbance? 2. Is the survival of the annuals at the scale of the whole grassland possible under the observed conditions of disturbance dynamics? 3. Which phases in the annuals' life cycle and patch types contribute most to population growth?
Location: Boreč hill, northern Czechia, 50°31′ N, 13°59′ E, 446 m a.s.l.
Methods: Local population dynamics of the annuals were analysed separately for five patch types that differed in the proportion of bare soil. Vitality rates were assessed directly in the field, but also in a garden experiment, during 2000–2001 and 2001–2002. Population dynamics at the scale of the whole grassland was analysed with a megamatrix approach, combining patch dynamics of the nest-mounds with patch-specific population dynamics. Contributions of different phases and patch types to growth rate were estimated by elasticity analysis.
Results: Nest-mounds differed in the percentage of bare soil. Increasing moss cover significantly reduced germination and seed production of all studied annuals and decreased their population growth rates (λ). Although successional processes dominated over ant disturbance, populations of all species could survive well (λ≫ 1) in the grassland according to the 2000–2001 megamatrix dynamics. Based on the dynamics from the following period, two species would not survive in a long-term perspective due to random environmental variation. Whereas the A-A transition (adult plants originating from adults of the previous year) had the highest elasticity under open conditions and 'good period' demography, the importance of persistent seeds increased under reverse conditions. This, however, differed among species.
Conclusions: Ant-disturbance was shown to be critical for the population survival of five annual species in the studied grassland. The fate of the annual populations in the grassland system also depends on random environmental variation, which may override the effect of ant activity.
