A direction finding ionosonde consisting of a seven‐element, 50‐m crossed baseline interferometer has been developed to provide group path delay, azimuth, and elevation of each resolved mode over the sounding frequency range. Direction finding (df) ionograms have been acquired automatically, using various noncooperative compatible sounder transmissions. A diurnal sequence of df ionograms obtained over a 2344‐km mid‐latitude path is reported, including plots of group delay, azimuth, and elevation versus frequency, as well as an azimuth/elevation scatterplot showing the distribution of angular occupancy for the path. The results show modes arriving closest to the great circle bearing (that is, with lowest azimuth error) arrive via the lowest elevation angles and with least group delay. Greatest bearing accuracy was observed on daytime sporadic E modes; measured at 0.85 MOF, the azimuth standard deviation was 0.3°. One‐hop F region low‐ray azimuth measured at 0.85 MOF showed standard deviations between 0.4° and 0.98°. Fractional degrees per megahertz linear azimuth gradients with frequency were also observed at various times of day. One‐hop F region high rays showed smoothly varying elevation as a function of frequency, increasing with group path delay as expected. Variable lateral deviation from the great circle plane was observed for high rays up to 5° in azimuth equatorward at various times of day; azimuth standard deviation at 0.85 MOF varied from 0.91° to 1.7°.

A direction‐finding ionosonde consisting of an interferometer using seven‐element, 50‐ and 150‐meter crossed baselines, has been developed to provide group path delay, azimuth, and elevation of each resolved mode over the sounding frequency range. Direction‐finding (DF) ionograms have been acquired automatically, using various noncooperative, compatible sounder transmissions. Diurnal sequences of DF ionograms were obtained for mid‐latitude paths at nominal ranges of 55‐, 1000‐, 2300‐ and 3500‐km. Data measured to characterize the DF properties of each path include group delay, azimuth, and elevation versus frequency, as well as an azimuth/elevation scatter plot showing the distribution of angular occupancy for the path. The results show modes arriving closest to the great circle bearing (i.e., with lowest azimuth error) arrive via the lowest elevation angles and with least group delay. Fractional degree bearing standard deviations were commonly observed for one‐hop E and one‐hop F low‐ray modes, decreasing with increasing range. Variable bearing standard deviations between 1° and 6° were observed for one‐hop F high‐ray and two‐hop F modes. Bearing standard deviation of 50° were seen at near vertical incidence (55‐km range).