In hydrological models, evaporation from interception is often disregarded, combined with transpiration, or taken as a fixed percentage of rainfall. In general interception is not considered to be a significant process in rainfall-runoff modelling. However, it appears that on average interception can amount to 20–50% of the precipitation. Therefore, knowledge about the process of interception is important. Traditional research on interception mainly focuses on canopy interception and almost completely denies forest floor interception, although this is an important mechanism that precedes infiltration or runoff. Forest floor interception consists partly of interception by dry soil, partly of interception by short vegetation (mosses, grasses and creeping vegetation) and partly of interception by litter. This research project concentrates on litter interception: to measure its quantities at point scale and subsequently to upscale it to that of a hydrotope. A special measuring device has been developed, which consists of a permeable upper basin filled with forest floor, and a watertight lower basin. Both are weighed continuously. The device has been tested in the Huewelerbach catchment (Luxembourg). The preliminary measuring results show that the device is working properly. For November 2004, evaporation from interception was calculated to be 14 mm of 42 mm throughfall (i.e., 34%).