Abstract
A procedure for the statistical analysis and the interpretation of factorial type field experiments for the assessment of fertilizer requirements is described and demonstrated using data from 49 field experiments in northern New South Wales. Yield data for each site are represented by the mathematical model : Y = b0 + b1N + b2N0.5 + b3P + b4P0.5 + b5NP + b6(NP) 0.5+ b7REPwhere N and P are nitrogen and phosphorus fertilizer rates respectively and REP =replicate. This model is used to solve the simultaneous nitrogen and phosphorus fertilizer requirements, using computer techniques. The data indicate widespread deficiencies of nitrogen and phosphorus, but, in general, only applications of phosphorus fertilizer can be justified at present prices. The development of more efficient procedures for applying nitrogen fertilizers or the reduction in their cost could lead to greater requirements for nitrogen fertilizers. The severity of the widespread nitrogen deficiency is, in general, not sufficient to seriously affect the estimation of phosphorus fertilizer requirements for the majority of sites.