Wild habitats adjoining farmland are potentially valuable sources of natural enemies, but also of pests. Here we tested the utility of birds as ‘sampling devices’, to identify the diversity of prey available to predators and particularly to screen for pests and natural enemies using natural ecosystems as refugia. Here we used PCR to amplify prey DNA from three sympatric songbirds foraging on small invertebrates in Phragmites reedbed ecosystems, namely the Reed Warbler (Acrocephalus scirpaceus), Sedge Warbler (Acrocephalus schoenobaenus) and Cetti's Warbler (Cettia cetti). A recently described general invertebrate primer pair was used for the first time to analyse diets. Amplicons were cloned and sequenced, then identified by reference to the Barcoding of Life Database and to our own sequences obtained from fresh invertebrates. Forty-five distinct prey DNA sequences were obtained from 11 faecal samples, of which 39 could be identified to species or genus. Targeting three warbler species ensured that species-specific differences in prey choice broadened the range of prey taken. Amongst the prey found in reedbeds were major pests (including the tomato moth Lacanobia oleracea) as well as many potentially valuable natural enemies including aphidophagous hoverflies and braconid wasps. Given the mobility of birds, this approach provides a practical way of sampling a whole habitat at once, providing growers with information on possible invasion by locally resident pests and the colonization potential of natural enemies from local natural habitats.

Reedbeds of the Rhône river delta, southern France, constitute an important breeding habitat for purple herons (Ardea purpurea) in Europe. However, the area of reed harvested commercially has increased in recent years and, since 1997, a European Union agri-environmental policy has been implemented to reconcile economic development with conservation goals for harvested reedbeds. Here, we investigate trends in breeding numbers of purple herons and the effect of reed harvesting on reedbed occupancy by herons. A regression model indicates that reedbed use by breeding herons was negatively affected by the reed harvesting pressure. An increase of nearly 50% in the harvested area following the implementation of the policy was probably caused by subsidies (from 67 to 100 US$/ha/year) proposed to reed harvesters, although associated with harvesting constraints intended to conserve heron colonies. The difference between the payments with and without colonies does not appear to have been sufficient to convince reed harvesters to conserve colonies. A case study showed that one heron colony located in a reserve area persisted, while another split into smaller colonies following modifications of reedbeds for future reed harvesting. Results suggest that the agri-environmental policy in its present form has failed to accomplish its primary objective of improving conditions for breeding purple herons in the Camargue. Potential breeding habitat for purple herons have decreased and urgent action is needed to protect some reedbeds from intense reed cutting.