Abstract
This study was aimed at examining the effects of reduction in the number of plants used for evaluation of the threshability of rice cultivars and its reproducibility. The four evaluators tested 50 plants, and grasped one, three or five panicles each time, using seven cultivars, three with hard, two with middle, one with easy, and one with very easy threshability. The effects of the number of plants used for investigation on the heritability of the shedding rate, which was estimated from the variance components of the two-way (cultivars and evaluators) analysis of variance, were analyzed 20 times by random sampling and the calculation of heritability on each condition. Heritability was reduced with decreasing the number of plants for investigation. The heritability was higher when the evaluator grasped three panicles than one or five panicles, and a very high heritability value was maintained even when the number of plants was reduced to 25 by grasping three panicles. Under this condition, it was possible to classify threshability into four grades, i.e., hard, middle, easy and very easy grade, but it was difficult to distinguish between very hard and semi-hard and also between semi-easy and easy threshability.
