Geiger-counter powder records of soda-rich plagioclases have been carefully measured and indexed. The lattice parameters of the synthetic specimens are almost independent of the lime content, the largest variation (that in β*) amounting to only 0·2°. After strong heating the powder patterns of the natural specimens closely approached those of the synthetic specimens.

Eight out of the nine natural plagioclases gave lattice parameters which fall within experimental error on a continuous line. There is a fairly large variation in lattice parameters, especially for γ*, which changes by 2°. As the lime content increases the lattice parameters of the natural plagioclases approach those for the high-temperature synthetic and heated natural specimens until at An₅₀ the separation is only one-fifth of that at An₀. The ninth specimen (from a dacite) gives parameters midway between the curves, thus exhibiting a state of partial inversion.

The partially heated natural plagioclases and the plagioclase from the dacite give parameters indistinguishable within experimental error from the parameters of unheated plagioclases with a higher lime content. There is, therefore, no reliable powder X-ray method for the determination of composition in the region An₀-An₅₀. If the composition is known the powder method may be used for the determination of the thermal state. If the plagioclase can be judged to be in the low-temperature state from geological evidence the powder method can be used to estimate the composition with an accuracy of 2 % An.