The zoning of plagioclase feldspars was studied with the four-axis universal stage. The composition, in terms of the extinction angle X’ʌ010, ⊥a, was measured for each zone in several crystals in each of nearly fifty rocks. The values of the extinction angles for each zone were plotted against the distance of the zone from the rim of the crystal, and the resulting zone-curves for each rock compared. . Three degrees of correlation between zone-curves are recognizable, and the rocks studied may be arbitrarily grouped on this basis as follows: 1. Rocks in which the correlation is poor or absent. These include lavas and some intrusive granites. 2. Rocks in which the correlation is good or perfect. These include hornfelses and, the foliated and gneissic metamorphic rocks. 3. Rocks in which the correlation is indefinite. This group, which contains only twelve percent of all the rocks studied, includes some intrusive granites and some foliated metamorphic rocks. The degree of mobility of magmas may be estimated. The mobility appears to be severely restricted by the continued growth of its feldspar phenocrysts, and by the confining walls of intruded country rock. The author concludes that extreme complexity and non-correlation of zoning are due to relative motion between growing crystals, and that these features are most common in magmatic rocks which have been thoroughly mixed during the fluid stage. He concludes further that simplicity and close correlation of zoning are due to the growth of crystals in a quiet, relatively undisturbed environment in which the crystals do not move relatively to one another. Finally, it is concluded that cyclic super saturation, as proposed by Hills, (1936), is a less important cause of plagioclase zoning than changing pressure-temperature environments due to mixing or turbulence.