time of the year, is to be expected more frequently during winter months with January, February, and March probably being the months of greatest expectation, The upper-level cyclone situation appears to have little, if any, dependence on season, Because of its per= sistence, up to ten days, it must always be taken into account in planning operations in the northern Marshalls, PART III Meteorology in Relation to Atomic Tests, It is probably safe to assume in planning future atomic tests in the Marshall Islands that each of the three categories of weather situation described in Part II will occur at least once during any period longer than one month. As already stated, there will be a tendency for low-level cyclones to be more frequent in summer and fall while the trade situation should be expected to occur with highest frequency in summer and spring. In the past, there has been a tendency to assume in planning more complicated operations that the wind systems aloft are associated only at random with ~ cloud and weather. Prior to Operation CASTLE, for exemple, there had been operational requirements that the winds over the Marshalls ‘up to 60,000 feet be, in the period following a detonation, from the southeast or south, At the same time, corollary air operations had been predicated on the assumption that trade cumulus without middle or upper cloud would prevail over the entire Marshall Islands area, Experience indicates that this is stipulating an incompati- ble distribution of weather elements. L7L It is true that such condi-