The effect of different N, P and K supply levels and their combinations on the development, air-dry matter, oil yield, and fatty acid composition of mustard (cv. Budakalaszi sarga, of Danish origin), and on the leaching and balance of NO3-N and SO4-S in the top 6 m of soil was studied in the 12(th) year of a long-term mineral fertilisation experiment set up on calcareous loamy chernozem. The ploughed layer contained approx. 3 humus, 5% CaCO3 and 20% clay. The soil was moderately well supplied with K, Mg, Mn and Cu and poorly with P and Zn. The experiment involved 4N x 4P x 4K=64 treatments in 2 replications on a total of 128 plots. The mineral fertilisers were applied in the form of 25% calcium ammonium nitrate, 18% superphosphate and 50% potassium chloride. The main results were as follows: 1. The development and yield of mustard was decisively influenced by the NxP supplies. As the result of N fertilisation the air-dry matter decreased by 2% in the rosette stage, 5% at flowering and 10% on average at maturity. P fertilisation reduced the dry matter content by 4-5% and K fertilisation by 1-2%. Ripening was protracted by as much as two weeks by over-fertilisation. 2. The air-dry mass of the shoots was 0.9 t (.) ha(-1) on May 23(rd) in the rosette stage, 2.7 t (.) ha(-1) on June 3(rd) at the beginning of flowering and 7.9 t (.) ha(-1) on July 15(th) at maturity. At harvest on August 22(nd) the stems had an average mass of 3.0 t (.) ha(-1), the pods 1.6 t (.) ha(-1) and the seeds 1.8 t (.) ha(-1), giving a total aboveground air-dry mass of 6.4 t (.) ha(-1). The approx. 1.0 t (.) ha(-1) seed yield on the control plots increased to 2.4 t (.) ha(-1) on plots given 140-150 mg (.) kg(-1) AL-P2O5 and 140-200 mg (.) kg(-1) AL-K2O, with an annual 200 kg/ha N fertiliser. On these same plots the total abovcground air-dry matter yield rose from 3.5 to 8.0 t (.) ha(-1). There was no reduction in yield as the result of over-fertilisation. 3. As the result of N fertilisation, which has a decisive influence in the generative phase, the thousand seed mass rose from 5.7 to 6.7 g, the oil content of the seed yield dropped from 31.5 to 27.9% and the oil yield rose from 370 to 590 kg(.)ha(-1). P fertilisation also significantly reduced the oil content by 1.3%, but the oil yield was increased by around 150 kg(.)ha(-1). After K fertilisation there was an average increase of 0.8% in the oil content and 10% in the oil yield. The 330 kg (.) ha(-1) oil yield recorded for the control plots rose to 700 kg(.)ha(-1) on soil given 140-150 mg (.) kg(-1) AL- P2O5 and 140-200 mg(.)kg(-1) AL-K2O, with an annual 100 kg (.) ha(-1) N fertiliser. 4. The mustard oil contained an average 41-42% crucic acid (C22:1), 22-23 % oleic acid (C 18:1), a total 19-21% of linolenic acid (C18:3) and cicosenic acid (C20:1), 10% linoleic acid (C18:2), 3.4% lignocerinic acid (C24) and 2.5 % palmitic acid (C16). After N fertilisation there was a significant rise in the percentage linoleic acid and lignoccrinic acid and a reduction in the linolenic+eicosenic acid %. 5. On soil with excessive N fertilisation NO3-N had been leached to a depth of 3.5-4.0 m by the 12(th) year of the experiment, representing a downward movement of around 30 cm (.) year, on this area, which has a negative water balance. The plant N uptake averaged 150 mg (.) ha(-1) (.) year(-1) on the N-fertilised plots. Some 30-50% of the non-absorbed N was detected in the soil profile in the form of NO3-N. The ratio of NO3-N in the soil increased with a rise in the level of over-fertilisation. 6. The maximum accumulation of KCl-soluble SO42- was found in the 1-2 m layer. Approximately 30% of the SO4-S introduced into the soil with superphosphatc and not taken up by the plants was detected in KCl-soluble form in the 6-m soil layer. The average moisture content of the soil layers decreased with depth as follows: 0-1 m 15%,1-2 m 13%,2-3 m 10%, 3-4 m 9%,4-5 m 7%,5-6 in 5%.
