Abstract:Some quantitative factors in agricultural experiment often have indeterminate levels. If this kinds of variation is ignored and the experiment result is analyzed by the traditional analysis of variance, not only the parameter estimates will have biases and the experiment error will increase, but als