In order to critical values of t to be precisely correct for testing Ho: u1-u2=0 using small independent samples, we must assume

a. ó1 = ó2
b. normality of populations
c. u1 = u2
d. all of the above.

D. All of the above.