After four years of consecutive applications of fertilizer mixtures containing P with or without S, in a long term trial conducted with cotton, on an acid and low-fertility Dusky Red Latosol, velvet bean (Stizolobium atterrimum Piper & Tracy) was cultivated and limestone enough to attain the soil pH around 6.2, was applied. During the next five years, from 1978 to 1983, IAC 18 cotton cultivar was grown with the same fertilization of the initial phase of study. N and K were supplied in constant dosis, and P and S in variable dosis, through the combination of ammonium sulfate, nitrochalk, ordinary superphosphate, triple superphosphate and potassium chloride. Sulfur deficiency symptoms such as the appearance of yellow leaves at the top of plants ("lemon green") occurred from fifth year on, in the plots which received no S, in accordance with a tendency of increasing yield. However, only after acidity correction, it was possible to maintain high cotton yield and to obtain differences statistically significant between fertilizer mixtures with or without S. In the last five-year period, annual application of 50 and 100 kg/ha of P2O5 increased cotton yield, respectivelly of 37 and 40%, as triple superphosphate and of 55 to 67%, in the case of ordinary superphosphate. With regard to economical aspects, boll weight and fiber length, ordinary superphosphate was also superior. The use of additional doses of S (120 kg/ha) however, did not improve the performance of cotton fertilized with 60 kg/ha of S. It was proposed that S should be incorporated in commercial fertilizer mixtures.
cotton; boll weight; fiber lenght ordinary and triple superphosphate