A new method is described that can be used conveniently in the field to determine the number of measurements of cross-strata dip directions necessary to obtain a significant average direction for an area. The method commonly used in the past has involved a great deal of graphical computation in the field. The suggested method is based on the fact that the number of measurements needed is approximately proportional to their variation. The estimated standard deviation of the first 50 measurements is determined, and, by simple linear regression equation, the number of measurements needed is computed. This regression equation has a standard error of plus or minus 18 measurements. If 36 measurements are taken in addition to the number indicated by the equation, the geologist can be about 95 percent confident he has enough measurements to compute an average dip direction that will be comparable to that which would be obtained with the standard cumulative vector direction curve method. With the suggested method confidence intervals for the average directions can be obtained easily. A brief discussion of the application of Stein's two-stage sampling method to cross-stratification studies is also given. Stein's method is applicable to the case where an average direction must be within previously specified confidence limits.