Sampling technique for estimating some quantitative characters of transplanted rice

Abstract

The study was undertaken to determine suitable sampling technique for the estimates of number of panicles and average panicle length/hill, number of filled and unfilled grains/panicle, grain weight/hill, grain weight/paniclc, and 1000 grain weight. A population of size 324 hills (18 rows and 18 columns) were considered and the sample of sizes 6, 9, 12, 18, 27, 36, 81, 108 and 162 were generated from this population. The coefficient of variation (CV) of the sample mean for each of these sample sizes were determined for both simple random sampling (SRS) and systematic sampling (SYS). From the comparison of SRS and SYS, it was observed that SRS is more precise than SYS. The sample size for estimating each of the above plant characters was computed by n=nQ/ {I+^Q/N)}, where n0 is a first approximation and N is the population size. The effect of population size on the sample size was examined by considering the population of size 36, 81, 108, 162 and 324 hills. The estimated sample size varied from character to character but for the same plant character, the sample size except for very small population, remained more or less the same. The sample sizes that seem to be adequate to have reliable (at 10% margin of error) estimates of the characters under study were determined to be 40 for no. of panicles and grain weight/hill, 25 for no. of unfilled garin/paniclc, 15 for no. of filled grain and grain weight/paniclc, 2 for average panicle length and 1000 grain weight.