W.L.LOW and C.Y. Tay 

Abstract

This study showed that spikelet number variation between panicles within hill is more important than between hills within plot or between plots. Sampling schemes involving the topmost three panicles within hill to evaluate this yield component are bias due to the large deviation from the actual average value for the hill. For practicality and reasonably good precision,a sampling scheme involving 5 hills and 5 panicles/hill is recommended for the evaluation of this yield component.

Full Text ( 286 KB )