Nitrogen leaching from the root zone in soils must be minimized to improve the chemistry of ground water. Nitrate leaching losses, even with optimized fertilization, are considered unavoidable. N leaching results in loss of N in the soil, which then must be replaced by fertilization; unavoidable N leaching indicates N surplus. The N balance can be used to characterize the potential hazard to ground water from agricultural use. The rate of unavoidable N leaching depends on soil hydraulic parameters and meteorological conditions. There is little information on the rate of unavoidable N leaching. Long-term lysimeter measurements are available for N leaching from profound Haplic Phaeozem soils in the Central German dry region. These measurements can be used to estimate the flow regime and unavoidable N-leaching rate. Results from lysimeter studies at a site on Haplic Phaeozem soil with silty-loamy texture developed from loess are as follows: water content at wilting point is between 0.167 and 0.270 cm³/cm³ and at field capacity (0.03 MPa) amounts between 0.286 and 0.342 cm³/cm³. The effective rooting depth is about 160 cm. The mean annual temperature is 8.2°C; the mean annual precipitation is 550 mm. In the years 1983 to 1994, field plants, like sugar beet, wheat, potatoes, barley and corn were cultivated. In the years 1995 to 2004, vegetable plants, like shrub bean, cucumber, cabbage, cauliflower, onion and peppermint grew on the lysimeters. The N balance was calculated from N supplied by fertilization, from seeds and legumes, and N removal by harvested crops. N deposition and denitrification are neglected in the N balance. The two lysimeters used in our studies are 2.5 m deep with a surface area of 2 m². They are monolithically filled and installed in an agricultural field to avoid oasis effects. To measure precipitation and evaporation they are weighed continuously with an accuracy of 0.05 mm. The percolating water is collected by ceramic suction cups with suction up to 0.03 MPa at a depth of 230 cm. In the years the lysimeters were planted with field plants (1983 to 1994) the annual average volume of seepage water was 11 mm. In four of twelve years there was no seepage water formation. Plants with deep rooting systems, like sugar beet and winter wheat depleted soil water by as much as 170 mm. Replacement of soil water takes place in winter, but only 108 mm are available on average. In the years the lysimeters were planted with vegetables (1995 to 2004), the amount of seepage water rose to 78 mm. N-leaching losses from field crop rotations amounts to 1.3 kg/ha and from vegetable rotations 19.4 kg/ha. The corresponding nitrate concentrations are 53 and 109 mg/l. The higher N-leaching losses from the vegetable rotation are the result of higher seepage rates resulting in double high nitrate concentration in the seepage water. Plotting nitrate concentration of seepage waters as a function of time shows a break-through curve during the period 1983 to 2004 at both lysimeters. The annual N balance varies from –70 to +34 kg/ha in the years 1984 to 1989. In 1990, N values peaked at +247 kg/ha, then decreased to –112 to –16 kg/ha in the years 1991 to 1994. In subsequent years, values ranged from –85 to 173 kg/ha. The nitrate peak of the break-through curve is affected by the high N-surplus balance in 1990. This result is supported by significantly smaller N balances before and after 1990. The nitrate peak was observed at different times with both lysimeters even though the quantity of seepage water was nearly identical (690 and 850 mm). This quantity of water corresponds to the water content of field capacity in the 230 cm soil column. Homogeneous matrix flow dominates in this soil. This indicates that N-leaching losses observed during vegetable cultivation result from displacement of soil solution enriched with nitrate over a longer arable land use period. The annual N-leaching loss during the period of study (1983 to 2004) was determined to be 7.8 kg/ha. This compares to an annual N balance of 9.5 kg/ha. The N input of deposition and the N output of denitrification seem to be of equal magnitude. In summary, N contamination of ground water descending from profound Haplic Phaeozem soils in the dry regions in Germany is small. Due to extremely slow water movement and homogeneous matrix flow, unavoidable N-leaching losses for arable land use are significantly smaller than 5 kg/ha/a.