average in stratospheric air than are found in the surface air below. The foregoing would seem to justify the use of surface air measurements as a means of determining, in a gross way at least, the time rate of depletion of the stratosphere as well as the mechanism and rate of transport via the stratosphere of debris from distant detonations. Measurement Techniques Air borne particulates were collected on HV-70 (Hollingsworth-Vose) 9 mil filters, 40 square inches in area, exposed in Suterbilt air pumps at flow rates of 25 to 30 cubic meters per hour. ulates of 0.25 micron and greater is 96 percent. The efficiency for partic- Weekly filters from a number of collectors located on the Argonne National Laboratory site and surrounding area were grouped by calendar month.° Each monthly group contained particulate matter collected from approximately 2 x 10° cubic meters of air. Approximately one-half of each monthly group (i.e., one-half of each filter) was partially ashed at 300 degrees C and compressed into a cylindrical disc for y-counting. Total ashing was not possible due to the asbestos base of these filters. Analysis for y-emitting radioactivity was done by Nal scintillation spectrometry using a 5-inch diameter, 4-inch-thick crystal in conjunction with a DuMont 6265 photomultiplier tube. Pulse analysis was implemented using an RCL (Radiation Counter Laboratory) Argonne Type 256 channel analyzer. Both sample and detector were placed inside of a 9-inch-thick steel shield, with an additional inner shield consisting of one inch of mercury inside a stainless steel annulus. Repeated spectrometric analyses of intact and ashed portions of monthly groups of filters showed no observable loss in activity of specific radionuclides due to the ashing process. At least two weeks elapsed between the time of collection and of counting a given batch of filters; thus any daughter products of radon or thoron would have largely decayed out. Cosmic-ray produced Be’, however, is still present in nearly its initial concentration after a two week delay. The method whereby the spectral data are qualitatively examined for the amount of activity due to specific radionuclides utilizes a standard reference source of known activity for each isotope present. The composite spectrum is then broken up into the appropri- ate number of energy regions, one for each isotope, and then the requisite number of simultaneous equations 103_2,,108 and cell ce l44, are solved. Due to spectral similarities, such as those existing between Ru use has been made of the radical difference in half life between members of these pairs. Hence repeated measure- ments were made of a given group of filters and the difference spectra analyzed for the short-lived components. 7 An example of this technique is illustrated in Figure 1 where the presence of Celt Be , and Zr°?-Nb?? in the difference spectrum is shown. The number andintensity of the individual spectral components changes drastically with time after collection or after atest series. Spectra of air borne fission debris collected 3, 15, and 25 months after the cessation of large scale weapons tests are shown in Figure 2. The decay of Zr 95. Nb 95 and the emergence of cst?” as the dominant radionuclide some two years after testing is clearly apparent. De- pending upon the circumstances, difference spectra and simultaneous equations are used, or a combination of simultaneous equations and successive approximations. cel4l- 144 Rut ??s 106° spt? csl37 An attempt has been made to routinely analyze for and Zre°-Nb°> in surface air filters. In addition, Be’, wi8h and Rn! 0? have also been assayed, the latter was frequently determined by chemical separation and subsequent Y-measurement,