Fine resolution measurements from the P78-1 satellite of 68- to 1120-keV electrons trapped at the lower L shell edge of the inner radiation belt have provided additional information on the previous discovery (Imhof and Smith, 1965) of the occurrences of peaks in the energy spectra. In some cases the data have yielded upper limits on the widths of the peaks as small as 26 keV. The energies of the peaks decrease slowly with increasing L value at a rate such that the calculated longitude drift periods remain nearly the same over the full L shell range for which they appear: often from ~1.2 to ~1.4 but occasionally up to L shells as high as ~2.2. These peaks are frequently observed at ~600-km altitude in the more durably trapped electron population having minimum drift altitudes of 100 km or greater and are to be contrasted with peaks in the quasi-trapped electrons precipitating from the inner belt at L?(1.5-1.85) that decrease in energy much more rapidly with increasing L value (Imhof et al., 1974, 1978; Vampola and Kuck, 1978). The spectra at the lower edge of the inner belt often contain sharp features which are continually changing, perhaps as a result of new injections and/or redistributions and the rapid loss of trapped electrons due to atmospheric scattering. For the narrow peaks the observed small but significant changes in central energy with L value that maintain a constant longitude drift period are consistent with the previous interpretation of the nearly monoenergetic electrons trapped on a given L shell resulting from a redistribution of electrons trapped on somewhat higher L shells (Imhof et al., 1966; Cladis, 1966). In that hypothesis the energy selectivity was postulated to result from a quasi-resonance process whereby the electrons were accelerated as a result of the geomagnetic fluctuations that had periods comparable to the azimuthal drift periods of the electrons. In one case presented here, narrow peaks were observed over a broad L shell interval on only one satellite pass but were not present 24 hours earlier or later, suggesting that the peaks might have been observed as a result of drift dispersion from an impulsive event.