The effects of different N, P and K supply levels and their combinations were examined on the amino acid content and yield of an established all-grass sward with seed mixture of eight grass species in the 28th year of a long term fertilization field experiment set up on a calcareous chernozem loamy soil. The lay-out and method of the trial, as well as the fertilizer responses on the hay yield and quality parameters, were published elsewhere (Kádár, 2005; Kádár and Győri, 2005). The soil of the growing site contained around 3% humus, 5% CaCO3, 20-22% clay in the ploughed layer and was originally moderately well supplied with available K, Mg, Mn and Cu and poorly supplied with P and Zn. The trial included 4Nx4Px4K=64 treatments in 2 replications, giving a total of 128 plots. The fertilizers applied were Ca-ammonium nitrate, superphosphate and potassium chloride. The groundwater table was at a depth of 13-15 m and the area was prone to drought. In 2001, however, the area had a satisfactory amount of 621mm precipitation with a fairly good distribution. The grass was established on 21. September 2000. The main results and conclusions can be summarised as follows:
     1. The content of GLU, ASP, HIS and ARG increased in the crude protein of the 1st cut hay as a function of N-fertilization up to 12-19%. The same time the content of PRO decreased up to 23%, CYS up to 25%, TRY up to 42% and ALA up to 48% compared to the N-control. The P fertilization raised the concentration of ASP, LEU, GLY, HIS, PHE and CYS, while the content of ALA dropped down to 41% that of P-contol.
     2. As a function of NxP positive interactions the content of ASP rose from 7.2 to 10.3%, HIS from 3.8 to 5.8%, ARG from 3.2 to 4.2% compared to the NP control, while the content of ALA diminished from 5.8 to 2.2%, TRY from 1.3 to 0.6%. The K fertilization also depressed the TRY synthesis. The N0P0K0 plots showed 1.59% TRY in protein, while the N3P3K3 maximum supply plots only 0.35%.
     3. The yield of amino acids was enhanced 3-4 times by N fertilization, 2.0-2.5 times by P fertilization and 20-30% by K fertilization. The amino acid yield increased on the N3P3K3 plots, compared to the N0P0K0 plots, in the case of ALA and TRY 3-4 times. Most of the other amino acids yielded 8-12 times more, HIS 14 times, PRO 16 times, GLY 18 times and CYS 20 times more. The maximum yield of essential amino acids made up 774 kg/ha, the total amino acid yield 1552 kg/ha and crude-protein 1779 kg/ha in the 1st cut hay.
     4. As a result of NxP interactions the ratio of ASP/ALA changed from 1.2 to 4.7, that of ARG/TRY from 2.5 to 6.3, that of HIS/TRY from 2.9 to 9.7 with the increased NP supply. Thus, fertilization can induce an imbalance of amino acids and so change or deteriorate the biological quality of protein. In similar circumstances the forage may require a TRY supplement to restore the imbalance of amino acids.