Summary: A classification of some clay minerals is made according to their lattice structure, and to the quality of their powder diagrams. The importance of the latter eriterion is emphasized.

The montmorillonite group is shown to have three end-members, montmorillonite, Al₂Si₄O₁₁, nontronite, Fe₂Si₄O₁₁, magnesium-beidellite, Mg₃Si₄O₁₁. The formulae given correspond to the completely dehydrated end-members, free from isomorphous replacements.

One montmorillonite, two beidellites, two nontronites, and one magnesium-beidellite are studied in detail. Their isomorphous replacements are calculated on the basis of eleven oxygen atoms and the excess cations are found to balance the negative charges resulting from the replacements. Similar calculations with the same result have previously been madeby Marshall. Base-exchange determinations on three of these materials show that all excess cations are exchangeable, but that in the case of magnesium-beidellite there are discrepancies, which require further investigation.

The hypothesis is made, that a certain amount of isomorphous replacement of silicon by aluminium and perhaps of aluminium by magnesium is essential for this structural type and may explain both the high water content, which is due to the excess cations, and the poor diffracting power for X-rays, which is due to the lack of regularity of the lattice.

Dehydration data given in the literature are discussed and the lack of agreement for the amount of hydroxyls given in the usual formula Al₂Si₄O₁₀(OH)₂ is pointed out.

X-ray powder data for the six samples and values for the lattice shrinkage upon dehydration are given. They show all samples to belong to the same structural type. The replacement of aluminium by iron and magnesium within the lattice causes an increase in the lattice dimensions and for air-dry material a decrease of the intensities of (003) and (005). Rough intensity calculations on the basis of Hofmann's montmorillonite structure (22) show corresponding values for (003) but too high values for (005), and the need for further experiments is pointed out.