Recursive Formulae for Constructing Block Designs with Dependent Errors
推导了在区组实验中增加新地块或观测值相依时更新C矩阵的递归公式，可用于从初始小设计逐步构建最优设计或序贯实验设计。
Recursive formulae are derived for updating the C matrix when extra plots in a block experiment become available, and when the observations are dependent. These formulae can be used in algorithms for optimal designs that grow from an initial small design, or for designs for efficient estimation of new contrasts of interest when an experiment is designed sequentially. Methods are also given for the efficient updating of the C matrix when treatments are exchanged or interchanged.