The importance of accurate knowledge concerning the causes of composition zoning in plagioclases lies in the possibility that such zoning may throw light on the magmatic history of the rock of which the zoned felspar is a component. The interpretation of plagioclase zoning has been the subject of much speculation among petrologists, Phemister having recently contributed (1934) a valuable discussion of earlier work, together with certain new suggestions as to the probable causes of the reverse and oscillatory types. My own attention was drawn to oscillatory zoning while studying the Upper Devonian igneous complex of Marysville, Victoria (Hills, 1932, pp. 152, 153), and I found the hypotheses suggested by Phemister and by Harloff (1927) to explain this enlightening. Nevertheless, it seems to me that the real significance of some of the processes postulated by both these workers has not been correctly indicated, and that the mechanism proposed by Phemister to account for oscillatory-normal zoning is not theoretically possible. Since it is important that the physico-chemical basis of such hypotheses be sound, a discussion of Phemister's and Harloff's theories, and certain other possible causes of reverse zoning is given below.