SummaryCoconut is harvested at bimonthly intervals (six picks per year) in all field experiments carried out by the Coconut Research Institute in Sri Lanka. Because recording of individual picks is costly and time consuming, it is of interest to know whether or not recording can be restricted to fewer picks without prejudicing the results of experiments. Yield data from five long-term field experiments conducted at five different locations showed that the sum of the first, third and fifth picks could provide a good estimate of the sum of six picks. Yield recording can therefore be confined to these three picks and the cost and time of recording in an experiment reduced by 50%. These three picks can conveniently be used to assess treatment effects on total annual yield.Análisis estadístico del rendimiento de coco
