Diagnosing sulfur deficiency in rice by grain analysis

The effect of varying the sulfur (S) and nitrogen (N) supply on the yield and composition of brown rice (Oryza sativa L.) grain was studied to determine whether grain analysis could be used for the diagnosis of S deficiency in this crop. Plants were grown to maturity in an S deficient yellow podzolic soil, under flooded and upland conditions in a glasshouse. In the first experiment there were six application rates of S combined with three application rates of N. The amount of grain per plant varied from 0.69 to 7.62 g, depending on the level of S and N supplied. Rice grown under flooded conditions produced approximately twice the grain yield of upland rice grown at field capacity. Grain from the flooded series generally had a lower S concentration, and apart from the grain from the low N treatment, had lower N concentrations than the grain from the upland series. The grain S concentration varied from 0.069% to 0.154% and the N concentration ranged from 1.06% to 2.14% with changing S and N supply. Strong positive relationships were obtained between grain yield and grain S concentration, and negative relationships between grain yield and the N:S ratios in the grain. The critical S concentration determined for 90% maximum yield in plants well supplied with N was not sufficient to distinguish between S responsive and unresponsive plants when N was limiting; both grain S concentration and N:S ratio are necessary to determine the S status of the rice crop. Deficiency was indicated when the S content of the grain was less than 0.1% and the N:S ratio was wider than 14:1, and the same criteria applied to rice grain from flooded and upland treatments. Note that these values were derived in a glasshouse experiment using one rice cultivar on one soil and are therefore preliminary and require confirmation before practical application in the field. The second experiment examined the effect of supplemental S applied after anthesis on grain composition. Late supplemental S had no effect on grain yield, or on the composition of grain from plants adequately supplied with S. In marked contrast, S concentration in grain of previously S deficient plants increased from 0.08% to 0.2%, well above the highest level achieved by applying S at sowing. It is concluded that grain analysis can be used to diagnose retrospectively S status for yield, provided S supply does not increase between the stage when grain yields are being determined and the subsequent grain filling stage. An increase in S supply between these two stages will change the relationship between grain S composition and yield and complicate interpretation of grain analysis for diagnosis. The advantages of using grain analysis for retrospective diagnosis of S deficiency are discussed, and the preliminary results suggest that the concepts warrant further testing in the field.