Results/Conclusions Production peaks in forest, wetland, and river ecosystems were partially offset seasonally. Wetland insect fluxes were highest in mid-July, with secondary peaks in mid-April and September, when river fluxes were lowest. Forest production peaked in early-June. The maximum river production occurred in early July as observed during previous years. In the wet season, insects were uniformly distributed among forest, river, and wetlands. Insect fluxes from forest to river and forest to river (averaged over 7 months) were similar (11 insects m-2d-1), but insect fluxes from wetland to forest were on average lower (8.5 insects m-2d-1) than from forest to wetland (13 insects m-2d-1). From late April through May, when the whole landscape was moist, there was no spatial variation in insect abundance-activity along forest, wetland, or river transects, and abundances averaged 315 insects m-2d-1. As the uplands dried out, from June to September, insect abundance remained high in wetlands and near the river, but dropped in the forest to average 32 insects m-2d-1. In contrast to the continental Hokkaido ecosystem, the peaks of river versus terrestrial insect fluxes were less offset in our Mediterranean system, and the seasonality of terrestrial versus river peaks was reversed. The wetlands, with three abundance peaks distributed through spring, summer, and fall, maintained insect fluxes when river and forest fluxes were low. Wetlands represent a miniscule proportion of this steep, seasonally dry landscape, but may supply critical food resources to insectivores during seasonal bottlenecks. We are investigating how insectivorous birds and bats track and respond to these seasonal landscape shifts in food supply sources.