The long-term 'biodegradation' on soil amino acids was examined in the control plots of '42 parcelles' experiment, established in 1928 at INRA, Versailles (France). None of the plots is cultivated, but is kept free of weeds, and mixed to a depth of 25 cm twice yearly. Topsoil (0-10 cm depth) samples collected in 1929, 1963 and 1997 were subjected to acid hydrolysis (6 N HCl) for comparison. The distribution and delta15N natural abundance of 20 individual amino acids in the soils were determined, using ion chromatography (IC) and gas chromatography-combustion-isotope ratio mass spectrometry (GC-C-IRMS). The total N and amino acid-N (AA-N), respectively, decreased by 54 % and 73 % in the period from 1929 to 1997. The average N loss was comparable for 1929-1963 (period 1) and 1963-1997 (period 2), but AA-N loss was three times faster in the former period. This significant reduction in total AA-N content was mirrored in the individual amino acids, which decreased by 74 % ± 1 % (ranging 58-89 %) between 1929 and 1997. The bulk delta15N values generally increased from 1929 to 1997, mainly associated with comparable or even higher increase of delta15N of the non-AA-N in the soil. The residence time (t1/2, time in which half of N was lost from a specific soil pool) was ca. 65 ± 5 years for the bulk soil, and comparable for periods 1 and 2. However, between periods 1 and 2 it decreased from 128 to 41 years in the non-AA pool, but increased from 59 to 92 years in the AA-N pool. Proline and amino acids that appear early in soil microbial metabolic pathways (e.g. glutamic acid, alanine, aspartic acid and valine) had relatively high delta15N values. Phenylalanine, threonine, glycine and leucine had relatively depleted delta15N values. The average delta15N value of the individual amino acids (IAAs) increased by 1delta unit from 1929 to 1997, associated with a similar rise from 1929 to 1963, and no change thereafter till 1997. However, the delta15N values of phenylalanine decreased by more than 7delta15N units between 1929 and 1997. The delta15N shift of IAAs from 1929 to 1963 and from 1929 to 1997 was not influenced by the relative amount of N remaining compared with the 1929 soil concentrations. The only exception was phenylalanine which showed decreasing delta15N associated with its decreasing concentration in the soil. We conclude therefore that in the absence of plant and fertiliser inputs, no change in the delta15N value of individual soil amino acids occurs, hence the original delta15N values are preserved and diagnostic information on past soil N (cycling) is retained. The exception was phenylalanine, its delta15N decreased with decreasing concentration from 1929 to 1997, hence it acted as a 'potential' marker for the land use changes (i.e. arable cropping to a fallow). The long term biological processing and reworking of residual amino acids resulted in a (partial) stabilisation in the soil, evidenced by reduced N loss and increased residence time of amino acid N during the period 1963-1997