The problem of sample size usually confronts the research worker in the initial stages of an investigation. Many short methods have been used for determining sample size such as extracting the square root of the number of boxes or taking a 10 percent sample. While these techniques are easy to use nevertheless they are not based upon a statistical evaluation of the experiment. Sample size in an experiment is often dictated by the amount of time, money and labor available but even under such conditions it is possible to evaluate the precision of the experiment when the number of replications is dictated by necessity.