Following two upward injections of energetic electrons (38 keV and 26 keV) from the Echo 4 rocket-borne electron accelerator (January 31, 1976, Poker Flat, Alaska), artificial auroral streaks were detected by ground-based low-light-level television. The streaks were recognized recently by using newly developed image-processing techniques. They were delayed relative to the injections by 2.06 and 2.42 s, respectively. These measurements of the bounce times of electrons of known energy constitute two accurate determinations of a high-latitude field line length (106¿1¿106 m). The delays are only 4-5% longer than calculated using a dynamic model of the geomagnetic field (Olson and Pftizer, 1982). Other field models yield shorter bounce times. Since the delays were in the inverse ratio of the relativistic velocities calculated for the nominal beam energies, it is concluded that the potential of the payload remained below 1 kV during 45 mA injections at an altitude of 210 km.

The echo streaks showed little dispersion in either time or space, indicating that the portion of the beam returing to the northern hemisphere loss cone remained collimated and nearly monoenergetic. But there was a 70% loss in the return flux. A diligent search failed to locate similar echoes from the more powerful injections employed in the Echo 5 and Echo 7 rocket experiments, suggesting flux losses of a least 98% and 92%, respectively. The losses are thought to be due to pitch angle scattering out of the loss cone as the electrons traverse the equatorial region but could also be due to collective beam plasma interactions. The lower loss rate in Echo 4 is likely related to the presence of only weak diffuse aurora during that flight. ¿ American Geophysical Union 1990