Air gasification of dried sewage sludge (DSS) in a fluidized bed has been studied as an effective alternative for the management of this material in a usable way. Nevertheless, one of the major issues in this technology is to deal with the tar formed during the process. To minimize the tar production, it is very important to optimize the operating conditions. In a previous work (Manyà et al, Energy Fuels, Vol. 19, 629-636, 2005) some unexpected results, in which tar production increases with the equivalence ratio, have been obtained. As it has been mentioned in that work, tar production could present unexpected trends while the bed composition changes from sand to a mixture of char and sand. Ashes contained in char have catalytic active metals which could promote gasification reactions. As the char content in the bed increases, the catalytic activity in the reactor could increase too, until the steady state is achieved. The aim of this work is to characterize the non-stationary period and its influence on the overall results obtained from gasification tests. Experiments have been carried out in a laboratory-scale BFB reactor at atmospheric pressure and at a reactor temperature of 1123 K with an equivalent ratio of 30%. Results show that at the beginning of the experiments the tar production was higher, until the steady state is reached. The gasification study has been enhanced with an analysis of tar composition by means of GC/MS and GC/FID. An effect of the transition period has been observed in tar composition also. The nitrogen aromatics percent increased with time whereas the polycyclic aromatic hydrocarbon decreased. Copyright © 2007 The Berkeley Electronic Press. All rights reserved.