3 CALIBRATION Since the energy of the gamma-ray emitted by Cs!*" (0.66 Mev) is less than that of the gamma-ray emitted by K*® (1.46 Mev), the spectrum of the radiation from the caesium is superimposed on part of that from the potassium. Determination of the Cs!” content requires that the contribution from the K** be known. The potassium content can be determined from the spectrum above the point where caesium contributes (about 0.8 Mev). A polythene phantom composed of right circular and elliptical cylinders was used for calibration purposes, It was filled with a strong solution of potassium chloride and the spectrum of the radiation was determined in the apparatus; by suitable arrangements of the parts of the phantom the effects of varying weight were determined, The spectrum was broken into four bands, the first two covering the caesium range and the second two covering the potassium range above 0.8 Mev. The sensitivity (expressed as counts per minute per gram potassium) in each band was plotted as a function of phantom weight. The sensitivity for any intermediate weight was determined by interpolation. The band limits andthe sensitivity in each band for a phan- tom weight of 70 kg are shown in Table 1, together with the background, determined with the phantom filled with distilled water. Table 1— BACKGROUND AND SENSITIVITY OF THE APPARATUS TO POTASSIUM AND CAESIUM IN A 70 KG PHANTOM Approximate band limits, kev I 115-525 Background, counts/min Counts/min/gram of K Counts/min/myc Cs'9? 261 i 525 — 820 76 0.382 0.0678 5.65 Ul 820 —1300 56 1.30 0.0807 IV 1300 — 1600 21 0.0569 Precisely similar measurements were made with the phantom filled with a weak solution of Cs'3? and the spectrum was divided into two bands (I and II). The sensitivity (counts per minute per mpc) for a phantom weight of 70 kg is also shown in Table 1. Two estimates of the potassium content of a subject were obtained from the observed counting rates in bands UI and IV, and the weighted mean was used to determine the contributions to the counting rates in bands I and II. These were subtracted from the observed counting rates and the differences used to calculate the caesium content. With the background known accurately, the statistical errors were such that the potassium content could be determined with a standard error of about +10 g, and the caesium content with a standard error of +0.7 muc in an observation lasting 50 min. For the measurements made using the single large crystal, the sensitivities to potassium in band IV and to caesium in band II were appreciably higher, while in the other bands they were Similar to those in Table 1. As a result the standard error (statistical only) on the potassium content was about +7 g, and on the caesium content +0.5 muc. 4 RESULTS AND DISCUSSION All the results obtained between June 1956, and July 1957, on subjects who have not been occupationally exposed to Cs'*" are set out in Table 2, With three exceptions (K.B., J.Be., and R.M.F.) these subjects are resident in Berkshire and Oxon. The results for subjects measured in June 1956 are slightly suspect due to variations in the background at that time. The mean potassium content of the 16 adult males is 0.212 + 0.005 per cent of body weight but with a standard deviation of +0.023 per cent. The large difference between this standard deviation and the standard error of a single observation (0.010 to 0.015), indicates that there is considerable biological variation. The value for the mean potassium content is consistent with the values obtained by ionisation chamber methods in 1953 and 1954 (Burch and Spiers, 1954; Rundo, 1955; Sievert, 1956)4* and with the value reported in 1955 by the use of gamma- ray spectrometry! (Miller and Marinelli, 1956); it is slightly higher than the most recent value 250 _ HARE