Conclusions concerning the unit-cell contents and chemical formula of a mineral are all too often drawn from a small part of the available data. A procedure is outlined by which any chemical analysis for which a density is available can be utilized, provided X-ray data are available for a reasonable range of analysed specimens. The effects of possible errors in the determination of water, and in the assessment of essential and non-exsential water, are discussed both qualitatively and quantitatively.

A survey of all available data for anthophyllite has not disclosed any specimens in which the number of oxygen atoms per unit cell is significantly in excess of 96 with the possible exception of the Glen Urquhart gedrite and the Edwards (New York) material. But it is certain that the number of cations per unit cell is normally well in excess of 60 (partial occupation of the A lattice positions), and that, although ‘excess’ water in sonic fibrous anthophyllites may be adsorbed impurity, the number of hydroxyl groups in others is well in excess of 8 per unit cell; it is also fairly clear that the number of hydroxyl groups may fall below 8 per unit cell.