A. E. Sharp, in Vowel-Length and Syllabicity in Kikuyu, examines one set of related orthographic questions and its phonologic background in detail. His objective is merely to determine ``what distinctions of length and syllabicity it may be desirable to make explicit in a Kikuyu orthography'' (59). To do so, he finds it necessary to examine the relevant parts of the phonology thoroughly and in detail. In the process, he develops some very significant observations about problems of a sort that are often difficult. A few of his examples are of very great interest, and the whole discussion of some importance for theory. His orthographic recommendations are no simplistic acceptance of phonemics on the one hand or of superficiality on the other. Rather he weighs each phonologic fact in the light of its orthographic usefulness. He concludes that some changes can made in the current orthography which will appreciably improve its usefulness, but hesitates to suggest precise graphic devices to effect these changes. I hope his suggestions are given the consideration they deserve in Kikuyu circles. This, however, will not exhaust their practical usefulness, as they rather clearly indicate what thorough phonologic investigation can contribute to orthography design. We need many more studies of this sort if the design of written languages is to be put on a sound basis.

One other paper deals with a phonologic problem: Vowel Harmony in Igbo, by J. Carnochan. This restates the already widely known facts in terms of prosodies. As a restatement it makes only a small contribution to knowledge of Igbo. But it would seem more intended as a tract advocating the prosodic theory than a paper directed to the specific problems of Igbo phonology. The paper has a certain value as a comparatively easy introduction to this approach, particularly since it treats a fairly simple and straightforward phenomenon where it is possible to compare it with a more traditional (though not structural) statement. It does show one feature of the system that has not been previously described. But it does not, as it claims, demonstrate that this could not be treated by traditional methods. It seems to me that it rather easily can.