ic elena na Seichica iiscnt SiTa ceeacy cece cata Sesha | the measurement reference, and tilt of the camera platform) depend upon the altitude of the plane and its distance from ground zero, the reference chosenfor the particular measurement, the measurements relating the several references, and the position of the bubble level showing platform tilt. The correctionsare applied to the film measurements with appropriate scale factors to derive the final readings of cloud diameter or height as functions of time. The data from the cameras are plotted for comparison and almost always show a gratifying agreement during the rise of the cloud to maximum altitude. This agreement showsthe analysis to have been correct. The diameterreadings and height to bottom always continue to agree well, but the height-to-top readings usually begin to spread apart after stabilization, disagreeing sometimes by as much as 20 percent. When discrepancies appear, the films and the measurement notes are re-examined to determinethe cause. Usuallyit isa matter of Judgment as to which portion of the image truly represents the top of the cloud. Sometimes it can be shown that the cloud must have moved enough to upset the scale factors; in such a case the average of two readings from opposite directions is better than either alone. A "best curve" is agreed upon, and estimates of error are derived from the scatter in the figures and from the knownuncertainties in the supporting data. 19 SECRET RT me eee aieem Kebenate ene rm enmee moo — mere On Mred are