STRATOSPHERIC DISTRIBUTION OF NUCLEAR DEBRIS 313 tests. The presence of high !9°Cd activities in the debris indicates that some, if not all, of the new “Sr had originated in the 1962 Dominic I rocket shots, particularly in the Starfish Prime shot, which produced the '%€d. By June 1963 the “Sr concentrations in the layer between 80,000 and 90,000 ft had increased and so had the ‘°°Cd concentrations. Some of the increase in Sr in this layer can thus be attributed to the downward movement of rocket-injected debris, but some may also have been caused by southward movement of debris from low-altitude in- jections in the northern hemisphere. By November 1963 Sr concentrations had increased in the layer between 55,000 and 70,000 ft, but they had decreased substantially in the layer between 80,000 and 105,000 ft. By May 1964 further increases in the concentration had occurred at 55,000 and 60,000 ft, but at all higher altitudes the concentrations had decreased, Can we attribute to the Dominic I rocket shots all the new debris found at 105,000 ft in December 1962? The nuclide ratios in the debris are generally consistent with such a hypothesis.° Did this debris actually move downward in the southern polar stratosphere during. 1963, leaving the higher altitude layers and reaching the 80,000- to 90,000-ft layer by June 1963 and the 60,000- to 80,000-ft layer by November 1963? The preliminary ‘°°Cd data reported by Salter® are consistent with this hypothesis. If these hypotheses are both correct, a mechanism for such movement must be found, andonly particle settling appears to be feasible. Turbulent exchange wouldnot produce movement opposite to the concentration gradient, and an organized circulation could not have moved air containing low *Sr concentrations into the upper region of the southern polar stratosphere because there was no stratospheric region from which such air could have been derived. STRATOSPHERIC BURDEN OF 2sr Using data from the balloon and the WU-2 sampling programs, we have estimated the distribution of *’Sr ina meridional plane through the atmosphere. These estimated distributions should adequately represent the actual distribution in the sampling corridor used by both programs, though extrapolation and interpolation of data into unsampled regions have been necessary. During the past few years, however, it has become clear that the sampling corridor has not always been completely representative of the atmosphere as a whole. This has been true during periods of weapons testing and during the first four or five months following the termination of such tests. On the other hand, we have reason to believe that within six months following a test series the injected debris is rather uniformly distributed in the zonal direction within the region of the stratosphere (tropical or polar) into whichit

Select target paragraph3