Dr. Lockhart: You can increase sensitivity of radioassays. most of those things. I think the limits that Les indicated were pretty good on One count per minute per sample is going to be just about the lowest limit for beta count- ing with a 10-percent standard deviation. You are assuming then that your chemistry and everything else is perfect; that is going to be a little hard to do, This is just about it on a statistical basis alone; there is not a lot you can do to improveit. Floor: I would agree generally with the numbers on activity requirements per sample that Les mentioned. feel that it is at all necessary to agree with respect to the area of sampler body. vertical distribution of material we are concerned with in these higher layers. I don't This is so sensitive to the Let's assume, just for the sake of reference, that we have one megacurie of activity in the high stratosphere, held up in a rather shallow layer. One megacurie would give you about 100 disintegrations per minute per 250 square centimeters integrated through the middle. This gives you 100 times the activity you need to measure Strontium 90 to 10 percent, using a discriminating low alpha-beta technique. Thus, you can see 10 kilocuries. Further, if you relax to 20 to 30 percent - in the case of Strontium 90, if you push present techniques, and if you put research effort toward this improvement proportionate to the importance of the cost of analyses - you might beat this by a factor of 4, be- cause you are dealing with quite hard data and you can make some other improvements. If you depended upon chemistry alone you could beat it by an order of magnitude, but if you want to discriminate by decay schemes, you are stuck in the range of 3 or 4 to 10 with the best possible improvement. I think one dpm is a good limit, therefore, for betas, and you would probably have to get to five, ten, or more dpm, depending on the decay scheme, for other things, particularly if you want to get a discriminating thing. You see, if we have a concen- trated layer in some region of the upper atmosphere we can easily pick up small fractions adequate for counting of 1 to 250 square centimeters. If things are not measurable at these levels, they are either of no consequence or our assumptions of distribution are erroneous. So we have to do this at more than one point in space before we can be sure. Col. Russell: I'd like to mention another possible assay method here: mass spectrometry. As you are probably aware, the mass spectrometer has reached a pretty high state of the art and is going to do pretty well at i101} gram samples. With regard to the question that Ed Martell proposed, of getting hold of the sample at 150,000 feet, or something like that, with the hope that one could detect an upper reservoir which may still contain significant quantities of some of the earlier very high yield thermonuclear tests, there have been some data published recently which give some information on the multiple neutron capture products resulting in very high flux gadgets. One which is quite accessible, for example, is Plutonium 242. About 50 percent of all the Plutonium 242 -- if my figures are correct -- produced to date was produced during our earlier thermonuclear test series. So that by examining the isotopic ratios of Plutonium 242, 241, 240, and 239, I think one might be able to assign a certain fraction of the debris to the earlier tests, say Castile, if any is still up there. Although this isn"t in the realm of high altitude discussions, I think, too, examination of ocean waters and ocean sediments for plutonium would lead to some rather interesting results and might, in fact, give us a little better balance than we now possess for total fissions we know we produced. 91