Fear in grasslands: the effect of Eurasian kestrels on skylark abundances.

Predation has received considerable theoretical and empirical support in population regulation. The effect of predators, however, could be achieved in direct (killing) or indirect effects (such as displacement). In this paper, we explored the relationship between Eurasian kestrels Falco tinnunculus and skylarks Alauda arvensis in Mediterranean grasslands. First, we analysed the presence of skylarks in the kestrel diet over 9 years. We also compared a grassland area of experimentally increased kestrel density and a second grassland as control area to evaluate the direct or indirect effect on skylark abundance. We also considered two different habitats, grazed and ungrazed plots. If skylark abundance decreased as the kestrel breeding season progressed in high-density kestrel area compared with the control area, it would suggest a direct effect (predator hypothesis). If skylark abundance remains constant in both areas of contrasting kestrel density, it would suggest that skylarks avoid kestrels (avoidance hypothesis). We found that skylark abundance decreased in the kestrel area from the beginning of kestrel nest-box installation to recent years. The rate of skylark consumption decreased in a 9-year period as kestrel abundance increased, although the total amount skylark consumption did not show a decreasing trend. In addition, skylarks were more abundant in the kestrel-free area than in the kestrel area. Finally, we found that skylark abundance did not change through the kestrel breeding period in relation to grazing. We suggest that an increased breeding density of kestrels during their breeding period may force the skylarks to breed in other areas, which may explain the decline of their abundance.