The effects of soil temperature and soil moisture content on the rate of loss of N‐(1‐ethylpropyl)‐2,6‐dinitro‐3,4‐xylidine (I, AC 92,553) were measured under controlled conditions. The time for 50% disappearance in a sandy loam soil at 75% of field capacity was inversely related to temperature (98 days at 30°; 409 days at 10°). At 25°, the half‐life increased with decreasing soil moisture content (122 days at 75% of field capacity; 563 days at 12.5%). In seven soils with different properties there was a trend towards a slower rate of loss as the organic matter content of the soils increased and the half‐life varied from 72 to 172 days, first‐order kinetics being obeyed. The herbicide was lost rapidly from an inert surface and 97% loss was recorded after 28 days at 25°. Losses from soil surfaces occurred more slowly and were greater from wet compared with dry soil. In the field, it was more persistent when incorporated than when applied to the soil surface. More than 60% of I incorporated in April 1975 could be detected the following September, but when applied to the soil surface, only about 20% of the applied dose remained by this time. Residues measured by gasliquid chromatography using a thermionic nitrogen detector closely paralleled those measured by a bioassay based on the root growth of buckwheat.