Archaeobotanical data from Late Neolithic lake-shore dwellings (4300-3500 cal B.C.) in the northern Pre-alpine lowlands are interpreted in different ways. The presence of permanent arable fields as well as arable fields with short fallow phases and shifting cultivation with slash-and-burn has been discussed. To test these hypotheses experimentally we have been conducting tests in a forest northeast of Stuttgart since 1994. The slightly south-exposed experimental area of approximately 4.5 ha on a loess soil is covered by mixed deciduous forest, is available for at least 20 years and has been divided into 34 plots of 30 × 30 m. Up to 2001, five plots were used for experiments. The normal procedure is clearing, burning the dry small timber (less than 10 cm in diameter) and then growing winter cereals (bread wheat) for one season. The yields were between 2000 and 4000 kg grains per ha. The harvest was more or less free from weeds. First attempts with summer crops gave much lower yields of about 1100 kg grains per ha. Continuous cereal growing on the same place in the following years resulted in minimal to zero yields, mainly due to vigorous weed growth. These weeds are not crop weeds, but forest perennials of clearings and forest fringes. Ploughing to remove the weeds is not possible, because of the presence of roots and tree-stumps, most of them still living. Weed regulation by hoeing, burning or cattle grazing remains to be tested. Protection of the crop from game and birds is by fences and nets, but protection from mice seems difficult. In the spectra from the pollen traps, clearing and burning are strongly indicated, but cereal growing only slightly.