For a reliable nutritional analysis, the most sensitive stage is the plant and soil samplings. The most proper sampling is the one that best represents the area under study, requiring a minimum of sampled plants and the least possible number of simple samples collected. This work aimed to assess the number of plants to be sampled for the diagnosis of nutritional status, as well as the number of simple samples for making up of the compost sample for evaluating the soil fertility in starfruit orchard. This study was performed in a commercial starfruit orchard, in Vista Alegre do Alto, State of São Paulo. Random sampling was taken, collecting the sixth leaf from the apex of the branch of 40 starfruit trees, at the medium canopy height during blooming season. In the same way, 30 simple soil samples were collected, in a zigzag pattern following the planting lines, using a Dutch-type auger, at depths of 0 to 0.2 m and 0.2 to 0.4 m. Assuming a sample error of 10%, 21 starfruit trees would be sufficient for determining the chemical leaf macronutrients. For micronutrients, a minimum of 52 sampled plants would be necessary. Increasing the number of soil samples resulted in reduced percentage of error within the desired average estimate, thus allowing for a recommendation of 14 to 17 simple samples at depths of 0 to 0.2 m and 0.2 to 0.4 m (error = 20%), respectively.
Averrhoa carambola; leaf sampling; soil sampling; mineral nutrition; starfruit; fruit tree
