INTRODUCTION This paper is intended: (1) to give an account of the relative seismicity of various parts of the earth during the limited period for which accurate information is available, and (2) to identify and discuss the geographical and geological relationships of the principal zones and areas of seismic activity. Maps purporting to show the distribution of earthquake epicenters have usually been based either on historical data or on the results of instrumental seismology, sometimes on both. Such maps, unless studied with critical attention, are likely to give a distorted impression of the facts. Historical macroseismic data are in general available only for land areas, and are much influenced by the present and past state of culture in the districts affected. Instrumental determinations require much careful sifting, for reasons which will become thoroughly apparent as the discussion proceeds. Many recent maps have been based on the epicenters published in the International Seismological Summary (Turner, et al., 1923–1940) and similar compilations of earlier date, without any attempt to discriminate with reference to the accuracy of the determination, the magnitude of the shock, and in most cases even the focal depth. Further confusion is created by alternative epicenters suggested for poorly recorded shocks. Thus, for the earthquake of September 19, 1926, at 20h,1 the Summary gives the following four alternatives: 42° S. 130° E., 72° N. 2.8° W., 47° N. 10° E., 59° N. 65° E. Probably none of these is correct; the shock appears to be a . . .