We have long known that the nominally anhydrous mantle minerals, such as can take substantial amounts of water into them (eg 3% for wadsleyite and ringwoodite). We also know that water has a drastic effect on the physical and chemical properties of these minerals, changing phase relations by hundreds of Kelvin and causing order-of-magnitude variations in viscosity, electrical conductivity and vacancy populations. This has created a whole genre of speculation about the water content in the mantle and its effects on mantle dynamics (eg; 'The mid-mantle water filter'; Bercovici and Karato 2003: 'Nine oceans of water in the transition zone'; Smyth et al 1994: 'No water in the transition zone'; Green at al 2010). We have, however, no firm understanding of the water content of the mantle deeper than about 100 km and these papers will remain speculative until we place firm constraints on the water content of the deep mantle. We can do this by combining geophysical inversions for electrical conductivity of the geomagnetic response function with laboratory measurements of the effect of water content on electrical conductivity. This is what the current project will do using cutting-edge designer-diamond anvil cell conductivity measurements with simultaneous infra-red measurements of water content. These new data will be used in Monte Carlo simulations of the geomagnetic response function which invert directly for water content.