采用计算机模拟的方法研究了NCⅠ、NCⅡ和完全双列杂交(DIA)设计在随交和自交群体及8种样本容量下估计遗传力和显性比的可靠性,并用变异系数和总体变量的容许区间两个指标对一次试验结果估计遗传参数的可靠性作了研究。结果表明:随着家系数增加,三种交配设计估计遗传力和显性比的可靠性都相应提高,但家系容量的增加只能在一定程度上提高显性比估计的可靠性。为了最大限度提高估计遗传参数的可靠性,应尽可能增加家系数,而家系容量只要为5到10就足够了;三种交配设计在任何情况下估计遗传力的可靠性均高于估计显性比的可靠性,特别对NCI和DIA设计,估计显性比的可靠性很低,除非家系数特别大。从一次试验结果估计遗传力和显性比的可靠性,即使在试验样本容量很大的情况下也较差,尤其是估计显性比的可靠性很差。因此应用一次试验结果估计遗传参数,必须重视它们的抽样误差大小。此外还对如何提高遗传参数估计可靠性的途径进行了讨论。 The reliability of estimated heritability and dominance ratio of NCⅠ, NCⅡ and DIA design with crossing and selfing populations and 8 sample sizes was studied by computer simulation. The variation coefficient and overall variable The allowable interval of two indicators of a test result to estimate the reliability of genetic parameters were studied. The results showed that with the increase of family coefficients, the reliability of the estimated hereditary and dominant ratios of three kinds of mating designs increased correspondingly, but the increase of family capacity could only improve the reliability of dominant ratio estimation to a certain extent. In order to maximize the reliability of the estimated genetic parameters, the coefficient of the family should be increased as much as possible, while the family capacity of 5 to 10 is enough. The reliability of the three mating designs in any case is higher than the estimated dominant The reliability of the ratio, especially for the NCI and DIA, is estimated to be very low reliability of the dominant ratio, unless the family coefficient is particularly large. Estimating the reliability of heritability and dominance ratio from one test result is poor even in the case of a large test sample volume, and in particular, the reliability of the estimated dominance ratio is poor. Therefore, the application of a test result to estimate genetic parameters, we must pay attention to their sampling error size. In addition, the ways to improve the reliability of genetic parameters are also discussed.