Mercury is a unique planet in many ways. In addition to its proximity to the Sun and the associated great heating, it has zero obliquity, with the Sun constantly on its equator and visible only as a crescent at the planet's poles. The 3/2 tidal resonance creates equatorial surface temperatures above 700 K. Nevertheless, it is pock marked with craters and there is considerable terrain within 5 degrees of both poles that is in perpetual shadow, creating cold fingers which can trap volatiles for perhaps billions of years. Because of the inclination of Mercury's orbit to the ecliptic, there are periods in each Mercury orbit that allow Earth observers to look partially down into these dark craters. This favorable circumstance occurred on April 8, 1991 when the Caltech radar astronomy group discovered the radar-bright feature on the north pole, which they interpreted as water ice in permanently shadowed terrain [ Slade, Butler and Muhleman, 1992; Butler, Muhleman and Slade, 1993]. The detection was immediately confirmed at Arecibo by Harmon and Slade [1992] and theoretical thermal model calculations by Paige, Wood and Vasavada [1992] provided a foundation for the ice interpretation. Subsequent work at the Very Large Array (VLA) by the Caltech group and at Arecibo demonstrated that the same phenomena occur at the south pole.
How can this be? There is no question that water vapor exists in the neighborhood of Mercury: from meteoritic impacts, the odd comet and, to some extent, direct outgassing from the planet's interior. The water molecules are very fragile in the presence of the solar UV flux with a half life on the order of 1 day. Nevertheless, Butler, Muhleman and Slade [1993] were able to show with Monte Carlo model calculations that a sufficient fraction of the water molecules will be captured and ``forever'' trapped on polar crater floors, behind suitably situated mountain ridges, and in cracks; sufficient to explain the observations [see also Butler, 1994 and Muhleman, Grossman and Slade, 1995 for more complete and updated discussion]. Recently, Harmon, et al. [1994] have demonstrated a strong correlation of the ice features with the set of polar craters imaged by Mariner 10. Many important questions arise, e.g., what secrets of the inner solar system are frozen in the deposits and how do we get samples of the stuff?