-17- While practical problems are central to this program, and basic meteorological research peripheral, the sampling cannot lift itself by its own bootstraps. When a body of data, such as the stratosphere or rain data, exhibits considerably greater variability or complexity than is consistent with the simple model on which the choices of sampling grids and intervals were based, it becomes doubtful whether an adequate sampling plan can be developed by studying the internal characteristics of this body of data alone. It’ is then pertinent to ask such questions as the following: (1) Are the variations due to organized phenomena on a smaller scale? (2) Are representative ranges of the controlling variables being sampled? (3) What is the likelihood of these variables combining to produce concentra- “tions of radioactivity at some time and place much greater than any that have been observed? (4) Are there large scale variations or trends which might to some extent invalidate material budget estimates or estimates of hazard based on the available data? The controlling variables are, of course, the same ones which control the global transport of heat, momentum, moisture and ozone. The AEC is supporting some basic research on stratospheric meteorology at MIT and at the Weather Bureau, employing these more familiar elements of synoptic and dynamic meteorology. For the foreseeable future, the AEC will continue to need quantitative estimates of atmospheric contamination and fallout under various actual or hypothetical conditions. Such estimates can seldom be made byOREARCHIVES... ing from data obtained by analysis of environmental samples. (7