AbstractPerennial ryegrass (Lolium perenne L.) was grown in pots on 21 UK soils, both with and without fertiliser N. The fertiliser N was applied in six equal applications of 15N‐labelled ammonium nitrate, each at the rate of 120 mg N per pot. The first application was mixed thoroughly with the soil, while subsequent applications were made in solution to the soil surface, after each of the first five of the six harvests of herbage. In the absence of fertiliser N, the proportion of the total soil N taken up by the plants, including stubble and roots at the sixth harvest, varied between 1.5 and 4.0%. In the presence of fertiliser N, the proportion varied between 2.1 and 4.7%. The apparent recovery of the fertiliser N was calculated from the difference between the amounts of N in the plants that received fertiliser N and in those that did not, expressed as a percentage of the amount applied. The actual recovery of the applied fertiliser N was determined by analysis of the plant material for 15N. With all soils at the first harvest, the apparent recovery was greater than the actual recovery. When calculated over all six harvests, apparent recovery of the total amount of fertiliser N was generally close to the actual recovery. This difference from the first harvest probably reflected (i) a reduction in the extent of turnover between fertiliser N and soil N when the fertiliser N was applied to the surface and (ii) a virtually complete uptake of available soil N by the end of the experiment, in both the absence and presence of fertiliser N. Differences between the 21 soils in actual recovery were not closely related, either positively or negatively, to a range of measured soil properties. A mean of 17.2% of the labelled fertiliser N was retained in the soil (excluding visible roots) at the end of the experiment. The lowest retention (6.2%) occurred with the soil which had the lowest contents of organic matter and silt plus clay but, with the other soils, the extent of retention varied only between 14.7 and 22.0% of that applied, and was not closely related to contents of total organic matter or macro‐organic matter, or to the C:N ratio of the whole soil or the macro‐organic matter.