Abstract
Fertilizer rate experiments were carried out on 24 sites in Gilan and on 25 sites in Mazandaran. On each site two replicates were laid down of a 32 NP factorial in which nitrogen (N) and phosphorus (P2O5) were applied at 0, 40 and 70 kg/ha in Gilan and 0, 45 and 90 kg/ha in Mazandaran. The main effects of N and P application on rice yield were related to soil characteristics by using multiple regression techniques. Relationships varied according to province. In Mazandaran province the much higher concentration of soil calcium carbonate negatively influenced response to and diminished the effectiveness of the fertilizer accordingly. In Gilan, where soil K status was moderate, fertilizer application diminished K uptake (as reflected by leaf analysis) and benefited the rice yield which showed a negative relation to leaf K concentration. Yield response to phosphate fertilizer in Gilan was negatively influenced by available soil P content, and response to 40 kg P2O5 per hectare diminished by 14·6 kg rice per hectare for every 1 p.p.m. increase in available soil P. It was shown how these results could be applied to the formulation of fertilizer recommendations. Post-harvest soil sampling of the experimental sites showed that application of phosphate fertilizer led to significant linear increases in available soil phosphorus but this decreased with nitrogen application.
