Nitrogen turnover in a loess catena

N-mineralization and mineral-N-contents were determined from spring 1993 to autumn 1995 on arable soils from loess (Luvisol, Calcaric Regosol, Gleyic- Calcaric Regosol and Cumulic Anthrosol) of a catchment area called "Bonarts hauser" farm near Gondelsheim in western Kraichgau (SW-Germany). The aim wa s to find out, whether the course of net-N-mineralization and mineral-N-con tent of the soil could be explained by the parameters sail temperature and soil moisture. Soil samples were incubated in polyethylene bags on site for mineralization and aliquots analyzed in the laboratory for the determinati on of the mineral-N-content. In 1993, 55 kg N ha(-1) were immobilized and u p to 170 kg N ha(-1) mineralized depending on the soil type. In 1994, betwe en 181 and 297, and in 1994, between 59 and 230 kg N ha(-1) were mineralize d annually. Mineral-N-contents of the different soils throughout the experi mental period (n = 45) were found to correlate much better (r(2) between 0. 55 and 0.86) whereas net-N-mineralization (n = 44) showed coefficients of d etermination (r(2)) just between 0.08 and 0.53. Except for the Luvisol (37% ) only 0 to 8% of net-N-mineralization could be explained by the combined e ffect of mean soil temperature and soil moisture at the beginning of the in cubation using multiple linear regression analysis. Merely 1 up to 9% of mi neral-N-content of the soil could be explained by the same effect. The vari ability and inconsistency shown by net-N-mineralization renders it inapprop riate as a tool for predicting nitrogen delivery of the soil and basis for setting rules governing permissible nitrogen amounts in the soil.