SIMBAD references

Exo-planets are preferentially found around high metallicity main sequence stars. We investigate whether evolved stars share this property, and its implications for planet formation. Statistical tools and the basic concepts of stellar evolution theory are applied to published results as well as our own radial velocity and chemical analyses of evolved stars. We show that the metal distributions of planet-hosting (P-H) dwarfs and giants are different, and that the latter do not favor metal-rich systems. Rather, these stars follow the same age-metalicity relation as the giants without planets in our sample. The straightforward explanation is to attribute the difference between dwarfs and giants to the much larger masses of giants' convective envelopes. If the metal excess on the main sequence is due to pollution, the effects of dilution explain why this is not observed in evolved stars. Although we cannot exclude other explanations, the lack of any preference for metal-rich systems among P-H giants could be a strong indication of the accretion of metal-rich material. We discuss further tests, as well as some predictions and consequences of this hypothesis.