Storm rainfall data from dense raingage networks in Illinois were employed in a study to determine the length of time required to obtain significance for various increases in storm rainfall due to weather modification efforts. The primary purpose was to evaluate the effect of stratifying the storm data on the detection of seeding effects for a given design using highly accurate measurements of the rainfall parameters. It was also desired to evaluate the efficiency of various rainfall parameters and the efficiency of various statistical designs in detecting various increases. Results indicate that the length of experimentation necessary for detection of seeding effects varies according to weather type, precipitation type, rainfall parameter, and statistical design employed. Results also indicate that as the seeding-induced increase becomes large, the choice of stratification, rainfall parameter, and statistical design becomes less important. An evaluation procedure is recommended which incorporates desirable features from several of the designs, stratifications and rainfall parameters considered in this study. Although it is difficult to verify, a 20% increase in precipitation can be detected in a five-year experiment provided proper choices are made of weather types, statistical designs, data stratifications and rainfall parameters.