在方差(变量)分析过程中。当F值达到显著水准时,为了进一步比较各个处理平均数差异的显著性,需要进行t测验。过去常用最小显著差数法(Least significantdifference,简写LSD),以一个共同的LSD值作为比较全试验各个平均数问差异显著性的标准。这在理论上是不够合理的,当试验处理超过5个时,可靠性就比较差。因此,1955年Duncan D.B.氏提出复全距测验法(Multiple range test),又称新复极差测验 In the variance (variable) analysis process. When the F value reached a significant level, in order to further compare the significance of the differences between the various treatments, t test is required. In the past, we used the least significant difference (LSD) method to use a common LSD value as a criterion to compare the significance of differences among all the means of the whole experiment. This is not reasonable in theory, and when the number of tests is more than five, the reliability is poor. Therefore, in 1955 Duncan D. B. proposed multiple range test (Multiple range test), also known as the new complex test