Variations in 0.2--3.2 MeV electron flux in the magnetosphere during the May 15, 1997, magnetic storm (the largest magnetic storm of 1997) are examined. After over an order of magnitude initial decrease of the 0.2--3.2 MeV electron fluxes, the 0.2--0.8 MeV electron flux at L0.4-MeV electrons from SAMPEX decreased at L>4.5 following the Dst drop but increased somewhat at L<4. GPS satellite data also show that the electron flux decreased in the energy range 0.2--3.2 MeV for all L values above the minimum detectable L value of ~4.2 simultaneously with the decrease in Dst, which is consistent with an adiabatic process. However, the recovery of the electron flux was different at different energies, with an earlier recovery of the less energetic electrons and a later recovery of the more energetic electrons. The recovery of the electron fluxes started before the recovery of Dst, indicating that nonadiabatic processes were involved. The 0.2- to 0.8-MeV electrons appeared in the low-L region (4.2--4.5) at about the same time that the GOES 9 spacecraft measured a strong depolarization of the Earth's stretched magnetic field. Outer zone electron fluxes continued to increase across a wide L range (L=3--8) though the electron flux exhibited a strong spatial gradient, with the peak flux below L=4.2 in the equatorial plane. These data are used to test the idea if radial transport from larger L can account for all of the increase in the flux in the heart of the outer zone electron radiation belt at L=4--5. However, the radial gradient of the phase space density for a given first adiabatic invariant was estimated to be negative as a function of radial distance during the time that the electron flux was increasing. This estimate is somewhat uncertain because of rapid temporal variations and sparse data. However, if this estimate is correct, the usual theory of radial transport from larger radial distances cannot account for all of the increase in the electron flux. The analysis thus suggests that another process, such as local heating, which does not conserve &mgr;, may be required to explain the subsequent enhancement of the more energetic (0.8- to 3.2-MeV) electrons but that additional data are required to answer this question definitely. ¿ 1999 American Geophysical Union