Solar abundance fits to the quiescent spectra 1-3 and 20-35 were made with a two-component thermal plasma (model 1a) whereas sub-solar abundance fits to the quiescent spectra were made with a one-component thermal plasma (model 2a). Modifications of these models were used for the flare spectra 4-19 in order to account for the `quiescent background'. Thus solar abundance fits to the flare spectra were made with a thermal plasma of two variable components plus two components kept constant at the average quiescent values T_qu,cool=2.4610^6K, EM_qu,cool=0.4910^53cm^-3, T_qu,hot =17.810^6K, EM_qu,hot=1.9510^53cm^-3, and z=1.00 (model 1b). Sub-solar abundance fits to the flare spectra were made with a plasma of one variable component plus one component kept constant at the average quiescent values T_qu=13.010^6K, EM_qu=5.2010^53cm^-3, and z=0.10 (model 2b). In both cases the mean values of spectra 3, and 20-35 were used to account for the quiescent emission.