Evaluation of plant density and yield interaction in maize production considering risk

In our polyfactorial 4-year long experiment, which was carried out on chernozem soil with lime deposits at the Experimental Farm of Debrecen University of Agricultural Sciences, we wanted to find out that out of four different plant densities (N1=60, N2=70, N3=80, N4=90 thousand plant ha(-1)) besides a fixed and most advantageous manuring level and tillage method, which one is the most favourable in irrigated and non irrigated treatment. Our aim was to choose the plant density by which a favourable yield can be achieved and the yield risk of which is acceptable for the decision maker. By using both mean (E) - variance (V) and stochastic dominance (SD) criteria we came to the same conclusion. In non-irrigated treatment N1 and N2 gave the best results for risk-averse decision makers but on the basis of neither criterion could we decide an order between the two plant densities. In the case of irrigated cultivation both criteria proved that the most advantageous plant density with regard to yield and its risk was N2 with 70 thousand plant ha(-1). With the E-V criterion we use only mean value and variance parameters to describe the effects of plant density, while with the SD criteria it is possible to make comparisons on the whole range of the crop. With smaller number of data the use of the E-V, while with larger data number the use of the SD criterion is suggested. We will get the same results from the use of the two criteria only if the factors and conditions of cultivation are well represented in our database.