Consider the situation described in Exercise 6 of Chapter 2 with t = 4 levels of the treatment factor, and
= 2:0.

(a) If by blocking the experimental units into blocks it was believed that the variance of the experimental error, 2, could be reduced by 50%, calculate the number of blocks that would be required to have a power of 0.90 for detecting a maximum dierence in treatment means as large as
.

(b) If by using a Latin-square design the variance of the experimental error,

2, could be reduced another 10%, determine the power for detecting treatment dierences when using a 4 × 4 Latin-square design.