The object of this investigation is to determine how far apart radio direction finders should be placed in order to obtain statistically independent samples, and to establish the number of radio direction finders to be included in the ensemble average. The conclusions are based upon high frequency radio direction finder bearing data observations obtained from two small aperture radio finders separated by a variable distance. Crossed-Adcock type direction finders were used and the separation distance was varied from 1/32 to 1/4 mi. The data were obtained over a period of approximately one yr during the time when conditions in the ionosphere were normal. (Author).