Consider a randomized complete block design with fixed block effects and fixed treatment effects. It is known that the a treatments in the design fall into two groups consisting of $a_1$ and $a_2$ treatments, respectively, with $a_1+a_2=a.$ Under the usual model and assumptions, it is decided to test if the average effect duc to the $a_1$ treatments in the first group is cqual to the average effect due to the $a_2$ treatments in the second group, against a two$-$sided alternative. Explain how you will test this hypothesis.