Many analyses of montmorillonite indicate liberation of water above 300°C in excess of the amount required by the structure proposed by Hofmann, Endell and Wilm (1933). The structure of Edelman and Favejee (1940) will account for a larger quantity of water, but other inconsistencies with experimental results seem to arise.

Using as an analogy the hydrogarnets (Flint et al.; Pabst ; McConnell), it is suggested that hydroxyl ions can occur as tetrahedral groups within the silica sheets in substitution for SiO4 groups. Liberation of this "tetrahedral water" is believed to be associated with the second high-temperature endotherm which is occasionally a small deflection on the D.T.A. curve but is, nevertheless, characteristic of montmorillonite.