4-5 GAS ANALYSES Although detectable quantities of gaseous radioactive isotopes of interest were measured in approximately all close-in gas samples, the results were disappointing, since there was extreme variation and inconsistencies in the proportionate concentrations of these isotopes throughout the samples obtained. Analysis of long-range gas samples were equally disappointing. In addition to the inconsistencies observed close-in, long-range samples were further complicated by lower concentrations of debris radioactive gases in the presence of significant background levels, particularly with respect to Kr®, which was to be used as a fission refer- ence. The variability of radioactive gas atom ratios for the close-in samples is probably due to unrepresentative samples of the cloud, as all samples were taken at altitudes well below the altitude attained by the main cloud. There is no guarantee that this variability would be eliminated by sampling at 75,000 to 80,000 feet for the megaton shots, however, as there are still insufficient data with respect to fractionation of the debris gases either with respect to each other or with respect to the particulate portion of the cloud. The environment of the explosion-~- water or barge shots at Eniwetok and Bikini---carry such large quantities of water into the atmosphere that serious effects in attempting to get representative and quantitative tritium measurements under these conditions would be expected, particularly during the first twelve hours after detonation when rainout and/or fallout is very prevalent. This factor appears significant, as the extreme variations in H*/Kr® ratios are normally not observed in sampling shots at the Nevada Test Site---e.g., as observed during Upshot-Knothole. This comparison is not absolute, since mgaton shots have never been fired at Nevada. However, during Operation Teapot, analysis of about ten shots gave gas atom ratios that were quite reason~ able within theoretical expectations. From this latter fact, it was _ concluded that the sempling equipment and laboratory analysis for the vari~ unreasonable the of sources close-in samples were not the principal ations observed in the gas data. The long~range samples may be compromised within the sampling equipment itself, since it is known that recovery of tritium from the sample containers, quantitatively, is open to serious question. Hence, the overall comparison of close-in analyses with long~ range analyses is not considered completely valid, because of the differ- ences in the sampling equipment used. It is anticipated that for Operation Redwing, sampling equipment will be completely converted to Squeegee type-~ both close-in and at long-range. thatl viewing the data on an overall and qualitative basis, it appears ratios for the Castle shots are significantly higher than those observed for Ivy~Mike. The range of values for Ivy~Mike is all Castle shots indicate ratios greater} “Majority of the samples analyzed. ratio _ Jin the Theoretically, based on reactions in 39