The first approximation analysis for deriving activation energy spectra from relaxation data, originally derived for linear distributed relaxation processes, is shown to be valid as well for nonlinear distributed processes. The analysis is also applied to nonlinear, undistributed processes and found to indicate a spectrum of activation energies with a maximum at the characteristic energy of the single process. The characteristics of spectra resulting from nonlinearities are considered, and a procedure for evaluating the possible physical significance of derived activation energy spectra is considered.