A correlation between foliar analysis and yield of corn was established, based on 31 fertilizer experiments conducted during the period of 1962-64, at 16 different localities of the State of São Paulo. In these trials five levels each of N, P2O5 and K2O were used; plots with and without sulfur were also included. Arranging in groups all the trials showing positive coefficients of correlation for total nitrogen (24 trials) and phosphorus (26 trials), the following correlations for total nitrogen (24 trials) and phosphorus (26 trials), the following correlations were found in each case: nitrogen, r = 0.51**, y - 4.26 + 28.66 x; phosphorus, r = 0.47**, y = 43.21 + 160.83 x. As the K content of levaves was affected mainly by soil type, the correlations for leaf potassium were obtained separately special consideration being given to this circumstance. The nitrate content of midribs, required for best yields, varies in the same year from locality to locality and in the same site, from year to year. This uncontrollable variability, of course, makes the interpretation of nitrate data difficult. The plants not receiving sulfur averaged low yields and had lower sulfate--sulfur content in the leaves. [The normal ones had 40 kg S per hectare but all plots received a central dose of 60 kg per hectare of each N or P2O5, and 30kg per hectare of K2O] . It seems that the mean values for calcium and magnesium in leaves, obtained from the range of levels which provides high yields, may be indicative of normal composition. Limiting leaf nutrient values associated with 80 per cent maximum yields in the field are as follows: nitrogen (N), 2.90%; phosphorus (P), 0.230%; potassium (K) - latosolic B terra roxa type of soil, 1.70%; gravelly podzolized soils, 2.80%; other soil types, 2.10- 2.70%.