4 Discussion

4.1 X-ray transient AGN

Why are there no "turn-ons'' among the sources of our sample? This is a selection effect due to the definition of our sample. Because our sample is chosen to include only sources that were bright during the RASS, we will be biased against sources that were faint then but have since brightened, but we include sources that have since become fainter - sometimes much fainter - than our original count rate limit. For obvious reasons, only the brightest sources were re-observed in later pointed observations. Fainter RASS sources can be seen only serendipitously brighter in pointed observations. Because of the much smaller coverage of the sky in the pointed observations (15% PSPC, 2% HRI; Voges, private communications) there is not much chance to detect a transient in its "high'' state. So far only one AGN has been reported being faint in the RASS and bright in a later pointed PSPC observation, RX J1242.6-1119 (Komossa & Greiner 1999).

What causes transience in AGN? The most common explanation of X-ray transience in AGN is a sudden increase of the accretion rate. This could be the result of either accretion disk instabilities, or even the tidal disruption of a star as has been suggested by Rees (1990). While in the Seyfert 2 galaxy IC 3599 both accretion disk instabilities or a tidal disruption of a star might cause the X-ray outburst (Brandt et al. 1995; Grupe et al. 1995a) in non active galaxies like RX J1624.9+7554, RX J 1242.6-1119, NGC 5905 (Komossa & Bade 1999) or RX J1420.4+5334 (Greiner et al. 2000) tidal disruption seems to be the more likely explanation. In the case of the X-ray transient WPVS007, Grupe et al. (1995b) discussed the possibility of a temperature change in the Comptonization layer above the disk as an alternative explanation. A lower temperature of this layer would shift the soft X-ray spectrum out of the ROSAT PSPC energy window (0.1-2.4 keV) and would mimic a dramatic change in the X-ray/Big Blue Bump flux.

Converting the HRI count rates of the two transients IC 3599 and WPVS007 into X-ray fluxes shows that they now have typical X-ray luminosities of a normal galaxy (about log $L_{\rm X}=33$ [W]). In both cases, we can consider this as their pre-outburst/transient luminosities.

4.2 X-ray variability

We find that low-luminosity objects have a higher probability of being found to be variable than the high-luminosity ones (Fig. 2). This result is not as prominent as in the samples of Boller et al. (1996) and Leighly (1999a,b). The reason is that their samples stretch over 7 and 5 orders in luminosity while ours only stretches over $\approx$2 orders. However, in our Principal Component Analysis on our original sample (Grupe et al. 1998a, 1999a) we found this result as a part of the Eigenvector 1 relationship. There are two explanations for the dependence of the variability on luminosity: a) the size of the Black Hole/AGN engine, and b) the number of the X-ray emitting regions (see discussion in Leighly 1999a). The $\chi ^2/\nu$ test for the short-term variability in the RASS data is a robust test. It does not take the length of the observation into account. However, the RASS coverages are usually in the order of days and therefore comparable. Only for a few sources the RASS observations were split into two parts half a year apart. We should mention that the result of the distribution of the variability seen in Fig. 2 is smeared out when the excess variance (e.g. Nandra et al. 1997; Leighly 1999a) is used instead if $\chi ^2/\nu$.

The comparison of the long-term variability (right panel of Fig. 2) is more complicated. The time gaps between RASS and pointed observation can vary from source to source between half a year up to 8 years. This is the reason why the HRI data points suggest less variability than the PSPC data, because they have been observed later than the PSPC. The only way to check out the long term variability of a large sample of AGN is repeated monitoring of the sky like it is performed for example by RXTE's All-Sky Monitor.

Figure 3 in Leighly 1999a shows that NLS1 are more variable than Broad-line Seyfert 1s. In our sample a large percentage are NLS1 (Grupe et al. 1999a; Grupe et al. 2001, in prep). NLS1 are also objects that have the steepest X-ray spectra (Boller et al. 1996; Grupe et al. 1998a). Therefore we checked for a relation between $\alpha _{\rm x}$  and the strength of the variability. This is what we find for the short-term variability throughout the sample: objects with steeper X-ray spectra (preferentially NLS1) show stronger variability than those with flatter X-ray spectra (see Fig. 3). This is in agreement with the findings of Green et al. (1993) where "sources with steeper energy spectra have higher normalized variability amplitudes''.

In principle, the same intrinsic processes that apply to X-ray transience (changes in the accretion rate or the disk temperature) also apply to the normal variability, but on a much lower level. In cases of very rapid variability, such as found in IRAS  13224-3809 (Boller et al. 1997), relativistic and Doppler boosting and gravitational lensing effects (see Boller et al. 1997; Leighly 1999a) have to be taken into account. The variability will be stronger amplified in steep X-ray sources.

Another alternative explanation of variability is a change in the cold and warm absorber column densities and their ionization states (e.g. Abrassart & Czerny 2000; Komossa & Meerschweinchen 2000).

We have shown that ROSAT with its All-Sky Survey and the later pointed observations was a well suited experiment to detect X-ray transient sources and to monitor the long-term behaviour of AGN. The best way to find more transients would be to perform all-sky surveys repeatedly.

Acknowledgements

We thank Drs Bev Wills, Mario Gliozzi, Wolfgang Voges, Stefanie Komossa, and Joachim Trümper for useful suggestions and discussions. We also want to thank our referee, Prof. Dr. A. Lawrence for his comments on the manuscript and valuable information on additional references. This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, Caltech, under contract with the National Aeronautics and Space Administration. The ROSAT project is supported by the Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie (BMBF) and the Max-Planck-Gesellschaft.

This paper can be retrieved via WWW: http://www.xray.mpe.mpg.de/dgrupe/research/refereed.html