Not all of the narrow line, type 2 Seyfert galaxies are associated to absorption along our line of sight. A fraction of the Seyfert galaxies have shown a strong decrease of their X-ray flux, on time scales of several years, that is not associated to an increased absorption but to an intrinsic drop of their activity (eg. Bassani et al. 1999b). One of the best studied cases is NGC2992 (Weaver et al. 1996). The fraction of these ``fossil'' AGNs is about 10% (Maiolino et al. in prep.). However, this is actually a lower limit to their real fraction since the identification of this class of objects requires X-ray observations at different epochs and, in particular, the source had to be in a high, bright state during the observations performed by the early, low sensitivity missions. These conditions are met only for a limited number of Seyferts. Obviously, the conservation of the number of active nuclei requires fossil AGNs to revive after a certain period, as it is actually observed (eg. Gilli et al. 2000a).
Shortly after that an AGN turns off, the BLR fades and the nucleus appears as a type 2 Seyfert whose X-ray emission is dominated by the cold reflection component due to the circumnuclear torus (Guainazzi et al. 1998). Although the AGN appears as a Compton thick Sy2, these features are not due to obscuration. About 10 years later also the echo from the torus should fade. If the nucleus remains in a quiescent state for an even longer period then also the gas in NLR clouds can recombine. Yet, the high ionization regions of the clouds should recombine much faster and, therefore, the observed narrow line spectrum should be characterized by low ionization lines similar to LINERs (Eracleous et al. 1995). In this phase the X-ray emission should be dominated by the non-variable warm reflection component with, possibly a highly ionized Fe line at 6.7-7 keV, as indeed observed in several LINERs. Some of the LINER nuclei might actually be fossil AGNs.