COMPARISON OF PROCEDURES FOR CONVERSION TO YIELD PER HECTARE IN THE EVALUATION OF TROPICAL VARIETIES OF MAIZE

Abstract

In agricultural trials, yield per hectare has been estimated from the area of the experimental unit, the average yield per plant, and the number of harvested plants. In yield trials of experimental maize (Zea mays L.) varieties, planted in several locations and growing seasons in the coastal zone of the State of Veracruz, four procedures were compared for the conversion of ear weight of the experimental unit to kilograms per hectare. The results showed that, based on lower error mean squares, lower coefficient of variation, and higher determination coefficient, the covariance procedure was the most adequate. The average weight per plant and the Iowa formula procedures, tended to overestimate yield; while, for better-performing varieties, by the covariance procedure, the adjusted yield was underestimated. A positive relationship between yield and number of plants per hectare was found for the application of the conversion factor (b=0.0910**), covariance (b=0.0136) and Iowa formula (b=0.0199) procedures; however, the relationship of the average weight per plant and yield was negative (b=- 0.0096). Covariance, with average weight per plant (r=0.88**) and the Iowa formula (r=0.98**), showed the highest degree of association.