We briefly describe the main features of the Brera Multi-Wavelet (BMW) survey of serendipitous X-ray clusters, based on the still unexploited ROSAT-HRI archival observations. Cluster candidates are selected from the general BMW catalogue of 20,000 sources based exclusively on their X-ray extension. Contrary to common wisdom, a clever selection of the HRI energy channels allows us to significantly reduce the background noise, thus greatly improving the ability to detect low surface-brightness sources as clusters. The resulting sample of ~250 candidates shows a very good sky coverage down to a flux \~3x10^-14 erg/s/cm^2 ([0.5-2.0] keV band), i.e comparable to existing PSPC-based deep survey, with a particularly interesting area of ~100 sq.deg. around fluxes ~10^-13 erg/s/cm^2, i.e. where highly-luminous, rare systems at z~0.6-1 can be detected. At the same time, the superior angular resolution of the instrument should avoid biases against intrinsically small systems, while easing the identification process (e.g. by spotting blends and AGN contaminants). While about 20% of the candidates are already identified with groups/clusters at z<0.3 on the DSS2 images, we have started a deep CCD imaging campaign to observe all sources associated to "blank fields". First results from these observations reveal a distant (z>0.5) bonafide cluster counterpart for ~80% of the targets.

3 pages, 2 figures; to appear in Proc. of the ESO/ECF/STSCI workshop on "Deep Fields", Garching Oct 2000, (Publ: Springer)