of a given object, H, is related to the size of the image, h, by the simple relation H = Dog h/f wheref is the effective focal length of the lens, and Dy,is the distance along the optic axis, from the lens to the plane (perpendicular to the optic axis) containing the object. The focal length is calibrated accurately for each lens. The imagesize and optic-axis distances, however, may both be poorly known. Assuming the distance to ground zero "is accurately known, it is not known how far away a given part of the cloud is (for example, the near edge), nor is it known how high that part is if ground zero or the horizon is obscured by clouds. Thus the gross cloud diameter may be determined more easily and reliably than the heights and positions of specified portions. It was to overcome these difficulties that use was madeof the stabilized platform. With this table the angle of tilt in aiming the camera is at least known, and film measurements can be made to fiducial or other artificial reference points where no natural reference appears in the photograph. Measurementsof this sort have been carried out successfully for objects lying near the ground-zero plane. To investigate such objects as the nearedge of the cloud, more information is required: for example, a second picture taken from another location. If the object can be clearly identified in both pictures and if the camera aiming angles are accuragely known, the position of the object can be computed by triangulation. This endeavor has failed for Castle, since the angles and positions are too poorly known and since two different views of most cloud irregu- . larities are quite difficult to identify positively. It appears that two views from nearly - the same position are needed, not more than 15 degree separation, and that the aiming angles must be measured with a higher order of precision than appears justified. 2.3 CORRECTIONS TO OBSERVED DATA The practical difficulties and limitations in the experiment are considerable, necessitating the evaluation of suitable corrections to the direct measurements. For.example, a correction must be made for the curvature of the earth and for refraction of the light path through the atmosphere. This is a relatively straight-forward calculation, involving the altitude of the plane and its distance from the object, and is treated thoroughly in many works (References 13, 14). It is employed whenever artificial reference points are used in measuring the film. A sirailar calculation is concerned with the position of the apparent horizon in the ground-zero plane. This horizon is used as the measurement reference whereverit is clearly visible. Since it usually lies behind the detonation point, it appears to be at some (easily calculable) altitude above ground zero. A more-serious correction is needed in determining the location of the top of the cloud, which may be obscured by other portions of the cloud closer to the camera. It is clearly impossible to photograph the top of a wide-spread objest unless the camera can be placed above it. This correction necessarily involves personal Judgment based on detailed examination of a series of pictures taken at different directions and times. A feeling is- developed for the probable shape at the top, and the data are sorrected accordingly. “ Additional corrections are> required to compensate for the variations in tilt of the stabilized platform. These variations are small, except when the plane banks and throws the table into the stops. Nevertheless the correction is significant even for the smal] variations and is considered as a separate step in the computations whenever artificial reference must be used. 13 SECRET ne mee veneer enntenn ome even we ee