E. C. Roelof, S. E. Hawkins III, D. K. Haggerty, and R. E. Gold
	Johns Hopkins University/Applied Physics Laboratory, Laurel, Maryland, USA

The onset times measured by ACE/EPAM of beam-like solar energetic electron events (38-315 keV) are not always consistent with simultaneous coronal injection at all energies onto interplanetary field lines. Only electron events in which the beam-like anisotropies endure into the event maximum phase are analyzed in this study. In this class of events, the intensity distributions have characteristic angular widths about the magnetic field direction that are comparable to the aperture full-angles of the detector collimators (48 deg). The actual pitch-angle distributions must have a smaller angular width. Consequently, the electron velocity at 1 AU parallel to the magnetic field is nearly equal to the full particle velocity (v). Because the center energies of the four electron channels cover almost a factor of two in velocity (0.39 < v/c < 0.73) or, equivalently, path transit times (21.2 min/AU > T > 11.4 min/AU), these inconsistencies in onset times at 1 AU correspond to statistically significant differences of several minutes in coronal injection times. When the peak intensity histories are displaced in time back to the sun assuming a parallel velocity of v, the lower energies are injected earlier. Although the length of the path from the Sun affects the absolute injection time, it cannot change the relative ordering of the injection times as a function of energy (assuming that the path length is independent of energy). We consider it unlikely that the path length is energy dependent, because the faster electrons would have to have the longer path lengths. This effect should be manifested at 1 AU in broader pitch-angle distributions at the higher energies, and we have not yet been able to establish such a systematic trend. On the other hand, if the path is energy-independent, then the dispersion in the injection times is real. In any stochastic acceleration process, the intensity must build up first at the lower energies before it can increase at the higher energies. Consistent with this picture is the independent observation that the intensity maximum (at the Sun) is reached simultaneously over our full range of energies. This time would then mark the end of the primary phase of the electron injection.