107. Suppose a level .05 test of H0: m1  m2  0 versus Ha: m1  m2 0 is to be performed, assuming s1  s2  10 and normality of both distributions, using equal sample sizes (m  n). Evaluate the probability of a type II error when m1m21 and n  25, 100, 2500, and 10,000. Can you think of real problems in which the difference m1  m2  1 has little practical signi cance? Would sample sizes of n10,000 be desirable in such problems?