Order and disorder - III. The structure of the intermediate plagioclase felspars

The lattice and fault system deduced in part II are considered in relation to the atomic structure, the known structure of anorthite serving as a starting point. Small amounts of Si above the 1:1 ratio are distributed randomly among half the Al$_2$ sites, causing faults with slip vector c$_0$ (where the subscript 0 indicates axes defined in part II). At x = $\frac{2}{9}$ the new structure is formed, with its excess Si concentrated in all B sites in two bridging subcells, the remaining 16 subcells having the (Si, Al) alternation of anorthite. With changing composition, Si continues to enter the same set of Al$_2$ sites, now causing faults on (100)$_0$ with slip vector -2b$_0$ as well as those with slip vector c$_0$, taken as on (010)$_0$, until this new set of sites is filled at x = $\frac{6}{9}$. Then a new kind of fault on (001)$_0$ appears, probably occurring randomly with slip vector b$_0$, and causing a progressive break-up of the (Si, Al) sequence of anorthite, hitherto preserved except in the bridging subcells. This fault accounts for the diffuseness and eventual disappearance of `e' maxima. There is some evidence of (Ca, Na) segregation on different A sites between x = $\frac{2}{9}$ and $\frac{4}{9}$. All observed diffraction effects in the low-temperature plagioclases are accounted for quantitatively, and the changes produced by heat treatment qualitatively. A qualitative discussion of the nature of the annealing process is included, with emphasis on the difference between the effect of misfits of a puckering nature and misfits due to atomic substitution. The implications of this type of approach for dealing with other materials showing subsidiary reflexions are considered.