ce ne ne eee cum ee ee ee ited by the number of significant figures in the least accurate value in the computation. One source of limitation in the posttest samples was the correction factor for decay, which was changing approximately 4 per cent per day and which was applied no closer than to the day the sample was counted. Another limitation occurred in the weight of the ash used to determine the correction for self-absorption. Samples were weighed to the nearest milligram, and aften the ashed weight was less than 100 mg and occasionally less than 10 mg. Also the correction factors for backscatter and for geometry were not determined more accurately than to two significant figures. 3.4 CORRECTION FACTORS 3.4.1 Geometry Geometry is about 50 per cent for an internal-gas-flow counting chamber. Hence the correction factor was approximately 2.0. 3.4.2 Backscatter Backscatter was previously determined for P™ tn our Nucleometers as being 30 per cent (see p. 24 of reference 1). The same correction factor, 100/130 = 0.77, was used again this year. 3.4.3 Coincidence Ld af The loss of counts due to coincidence was determined empirically. Small amounts of P® were dried on tinfoil, and the pieces of the P* tinfoil for which the counting rates had besn determined were placed one at a time side by side on the counting plate. After each piece of the P*tinfoil had been placed onthe plate, a count was made. The expected count, which was the sum of the individual counts, was divided by the observed count to determine the correction factor. For observed counts less than 80,000 c/m, no correction was made for coincidence; for counts of 80,000 to 160,000 c/m, the correction factor was 1.01; for counts of 382,000 to 392,000 c/m, the correction factor was 1.07. Correction factors were also determined for in- termediate values. For small corrections the counting rate, as determined theoretically from =. the formula N = n/(1i — nt), held true (N = true counting rate, n = observed counting rate, and t = recovery time of register, or 5 psec). 3.4.4 Background The background values were determined by interpolating between background counts at the were taken during a 24-hr period. From November 24 to the end of December, 127 background counts were made with counter 185, and the mean value was 50.25 c/m, with a standard deviation of 2.63; for counter 184, 135 background counts were made during the same period, with a mean value of 53.51 c/m and a standard deviation of 3.84. These background values are less than those recorded in the 1949 report because the counting chambers are now shielded with 2 in. of lead. 3.4.5 Decay At the time of counting, the posttest samples were decaying at an appreciable rate. Therefore these samples were corrected for decay, and the date to which they were corrected was arbitrarily chosen as December 1, one month after Mike test and also near the mid-point of the counting period for these samples. Although the activity of the samples is corrected to that of December 1, the distribution of activity is that of the date of collection, November 2 to UNCLASSIFIED | REGERIGLE 9 ~ point (time) when the sample count was made. Usually about five 20-min background counts eke