Resumen
- Evaluations with a uniform plant stand are fundamental for efficient selections in cassava (Manihot esculenta Crantz) breeding. However, it is difficult to correct data of missing plants. The overall objective of this study was to analyze yield losses in agronomic trials and to develop a function that satisfactorily adjusts plot yields affected by a varying number of missing plants. Eight clones were initially evaluated in different locations for five consecutive years. For every genotype, mean plot yields decreased as the number of missing plants increased. Average losses ranged from 10.6 to 78.8% by removing from one to eight plants, respectively. Yields per plant increased significantly when more than four plants were removed due to a compensatory growth effect. Graphic analyses showed that the power function was the best explaining the relationship between fresh root yield and number of harvested plants. A model that properly adjusted yield for all varieties but one is proposed. Hopefully, this formula to adjust yields will help to improve the quality of cassava trials worldwide.