The effect of different N, P and K supply levels and their combinations were examined in the 33rd year of a long-term fertilization experiment on the yield and mineral element content of a 6 years old established all-grass sward in 2006, with seed mixture of eight grass species. The trial was established on a calcareous chernozem soil. 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 4N×4P×4K=64 treatments in 2 replications, giving a total of 128 plots. The fertilizers applied were Ca-ammonium nitrate, super phosphate and potassium chloride. The groundwater table was at a depth of 13-15 m and the area was prone to drought. The 1st cut was made on 08th June, the 2nd one on 11th September. During the vegetation period of 8.5 months in 2006, the site had a total of 397 mm precipitation. The lay-out, method and main results of the trial were published earlier (Kádár, 2004, 2008; Kádár és Győri, 2004, 2005). Main conclusions of this study are as follows:

1. The 1st cut hay yield gave the ¾ of the total yield. Highest yields were reached with the 200 kg/ha/year N-fertilization on soil well supplied with P (Ammonium-lactate soluble P2O5: 214 mg/kg). The yield of NP control plots increased from 1.5 t/ha to about 7.5 t/ha as a function of the N×P positive interaction. The rising P supply alone was not able to enhance the yield, however the N fertilization gave 3.5 t/ha hay surplus even in the P-control treatments. N and P fertilization together resulted in 6.0-6.5 t/ha yield surpluses. The effect of K fertilization also reached 1 t/ha on the NP levels where the ammonium-lactate soluble K2O content fell below 150 mg/kg.
2. The 2nd cut hay gave 0.5 t/ha on the NP-control plots unfertilized for 33 years, and 2.5 t/ha on the 300 kg/ha/year N treatment with well P-supply. The two cuts together resulted in yield levels between 2-10 t/ha according to the extreme NPK supply. In this year, with relatively good amount of precipitation, the hay yield surpluses for 1 kg N were 47-33-23 kg for the 100-200-300 kg/ha treatments.
3. The C/N ratio of the 1st cut was narrowed (from 52 to 24) with N-supply and the concentration of N as well as most of the cations increased with the rising N fertilization. In the low yield of the 2nd cut the elements, metals were accumulated. The P, S and Sr were enriched in hay as a result of rising P supply, as superphosphate contains these elements. Antagonistic effect of P predominated in the uptake of other elements, metals (Na, Zn, Cu, Mo, Cr, Co). K content of the hay was lifting while other elements were dropping with the increasing K fertilization partly as a result of dilution effect (N, P, S) and mainly because of cation antagonism (Ca, Mg, Na, Sr). K-B antagonism also appeared.
4. The N×K interactions resulted in 2-fold Sr and 18-22 fold Na content changes while N×P caused 18-22 fold changes in Mo contents, especially at the 2nd cut. As it can be seen, fertilization can have drastical effects on soil and crops. The induced element deficiencies or oversupplies can lead to diseases, disturbances in the metabolism of animals, so the soil and fodder analyses are necessary.
5. Considering the leaf diagnostical data, the satisfactory level will be at 200 kg/ha/year N supply and 150 mg/kg ammonium-lactate soluble P2O5 and K2O level or above. The S, Ca, Mg, Fe, Mn supply were satisfactory even at the control plots, while the Zn, Cu and B levels showed deficiency. The P/Zn and K/B ratios became adversely wider in some treatments, as well as the narrowing of the Cu/Mo ratio denotes Cu deficiency and Mo oversupply.
6. The amount of elements uptaken by hay as a sum of the two cuts and as a function of the supply/yield varied between the following values in kg/ha: 17-163 N; 36-122 K; 9-48 Ca; 6-17 P; 4-15 S; 3-14 Mg; 0,3-8,0 Na; 0,2-1,4 Fe; 0,2-0,9 Al and Mn. The other elements showed the following uptake: Zn 33-194, Sr 28-141, Ba 5-46, Cu 5-39, B 5-26, Mo 3-6 g/ha.
7. The botanical composition was drastically modified by the aging of the grass and the nutrient supply. Only three species remained out of the eight sown species and one immigrated. Coverage of the tall fescue was between 21-70% according to the N×P supply and 44% as average; coverage of cocksfoot varied between 4-24% depending on the treatment and 18% as an average; coverage of crested wheatgrass was between 0-28% and 9% as an average; the immigrant smooth brome covered 0-24% and 9% as an average; Weed cover was 3-4% as an average at the 1st cut. Weeds thrived mainly on those areas where the grass thinned away (extreme NP-deficiency or oversupply). The total plant coverage on NP-deficient soil was about 50%, while on treatments well supplied with NP it amounted 95-97%.