In the Central Pacific, especially to the eastward of the 180th meridian, it is very difficult to find any very reliable traces of this supposed universal seasonal variation in available data. It is true that there is a great variability in weather, but it seems to occur with little relation to the time of the year, In fact, the re. gion overlapping the equator and covering the longitudes of the Ha-~ waiian chain is a zone with probably the most highly variable rain- fall of the entire globe, The Marshall Islands further to the south. west fall remotely under the Australian influence but, as a consequence, witness only slight seasonal variations in cloud and rainfall. And even this influence is overshadowed by a greater, aperiod- ic variability of the type found further east in the Central Pacific, Even to the south of the Marshalls this variation extends to the winds. For example, at Ocean Island lying near the equator and south of the Marshalls, west winds sometimes replace the more usual east winds for several days at a time. This change is only vaguely con- nected with season and it is incalculable in its occurrence from year to year. Records maintained by the British since 1900 show that there have been years in which westerlies prevailed at Ocean Island on as few as two days and as many as 167 days. At such a location, then, one could not conceivably depend upon the seasonal occurrence of west winds in any given year nor could one plan military opera- tions there on the basis of statistical presumptions with any reasonable assurance, This, then, is an indication of the sum and magnitude of the meteorological problem confronting successful atom- 160