A standard method for the determination of glucose in serum is reported to have a standard deviation of 0.40 mg/dL. If s = 0.40 is a good approximation of lower case sigma, how many replicate determinations should be made for the mean for the analysis of a sample to be within: a. 0.3 mg/dL of the mean 95% of the time b. 0.2 mg/dL of the mean 95% of the time