For the stake locations previously discussed, there were no changes in the radiological conditions at these sites between the two sets of measurements, so far as is known. Individual stake locations have been remeasured on 20 or more oceasions because (1} fine grid data were required where previously a coarse grid had been measured, or (2) the validity of a measurement was doubted. These "reproducibility tests" were not formally analyzed; however, in many cases repeat measurements were within 10 percent of the first measurement and in most cases within 15 percent. Should a more exact value be desired for the overall reproducibility of IMP measurements, a formal study of these repeats is reeommended. TABLE B-21-3. IMP REPRODUCIBILITY STUDY Island: PEARL Stake July 1978 Detector 496 3-N-2 17.3 16.2 1.07 1-BL-0 14.6 12.2 1,20 9-S-3 21.9 18.4 1.19 6.9 7.0 0.99 3-BL-0 March 1979 Detector 396 Mean Ratio Ll] + 0.10 Island: LUCY Stake March 1978 Detector 496 March 1979 Detector 396 Ratio 10-W-8 2.3 2.9 10-W-6 12.9 12.1 1,06 10-W-4 21.1 19.8 1.07 10-W-2 21.5 21.0 1,02 10-BL-0 19.7 22.5 0.88 0.8 0.97 + 0.12 Both Combined Mean 103 + 0.13 ~7 Mean B-21-5