Preliminary results are presented from an empirical study of sherds from six middens investigated during the 1975 Draper Site rescue excavations. The results suggest that the current practice of arbitrary selections of sampling designs and sampling fractions are inappropriate, and recommendations for their choice are presented.