The Performance of DRIS and critical value approach (CVA) diagnostics were evaluated based on leaf analysis from a field experiment on a factorial design to study the yield response of banana to N and K applications during two crop seasons. The situations identified as deficient were classified as positives (+) and those diagnosed as sufficient, were classified as negatives (-). According to the response on yield due the application of N or K, the diagnostics were classified as true (V) or false (F). Four combinations were possible: V+, V-, F+ and F-. The efficiency for N diagnostics (%V+ and %V-) based on the critical value approach was 48%, while using the DRIS it was 69%. The efficiency for K diagnostics was 63% for both DRIS and CV methods. The relation between deficient diagnostics confirmed by positive response on yield under application of the nutrient and false positive cases (%V+¸%F+) for N was 1.5 for the DRIS and 0.68 for CVA. For the K deficiency diagnostics, the value of this relation was 1.67 for both methods. The net yield derived from N application based on correct diagnostics (48 cases) was 124 Mg ha-1 for DRIS and 20 Mg ha-1 for the CVA. For K, this variation was 70 Mg ha-1 for the DRIS and the CVA.
Musa spp.; plant analysis; critical level; mineral nutrition; nitrogen; potassium