78 D. CHAKRAVARTI AND E. E. HELD appear to be completely out of line. Repetition of the analysis of these samples gave essentially the same results and no explanation for the disparity is offered. A significant difference in potassium levels associated either with island or season of collection is not apparent (March falls in the dry season and August in the wet season). Nor is there any correlation between potassium and cesium-137 levels at any one island. In this regard, the data should not be pooled since it has beenestablished that there are differences in cesium-137 levels from island to island. The regression of ceisum-137 on potassium in the crab muscle, as determined in the largest sample from Kabelle Island, indicated no relation between these elements. Cesium-137 levels have been reported as cesiwin units, micro micro-curies of cesium-137 per gram of potassium (U.S. AEC reports on fallout). This unit is based on metabolic similarity of cesium and potassium andis used to facilitate comparison of different types of materials in a mannersimilar to the use of the strontium unit adopted by Libby (1956). Anderson, Schuch, Fisher and Langham (1957) have used the gamma ratios of ceisum-137 to potassium-40 in place of the cesium unit. Cesium units for Birgus muscle are given in Table 1 for comparison with other published data on fallout given in the same units. However, since the potasium and cesium levels are-both highly variable in Birgus muscle, there 1s some doubt as to the usefulness of the unit in this case. For example, the comparatively high value for cesium units of samples from Rongelap Island is due to the exceptionally low potassium content of one of the samples. On the other hand, the high value for the March 1959 collection at Kabelle Island is due to high levels of cesium-137, while the potassium levels are about average. These differences are not surprising when the following facts are considered. Gross (1959) has reported that Coenobita, which has habits similar to those of Birgus, obtains blood salts from its food. From work in progress at this Laboratory it is known that Pandanus fruit, which form a part of the diet of Birgus, vary by as much as a factor of five in cesium-137 content and a factor of four in potassium content and that the range in potassium levels in soils on which Pandanus growsat Rongelap is 0.1 to 1.5 milliequivalents per 100 grams. Also, the extremes in the range of gross beta activity in the top inch of soil at a single island can differ by more than a factor of ten. Ona basis of the variability in cesium-137 and potassium levels found in the environment alone, one might expect the high degree of variability found in Birgus. An additional factor which may accentuate the condition found in Birgus is currently under study at this Laboratory in cooperation with the Department of Botany, University of Washington. The low potassium levels in the soils at Rongelap represent a potassium deficiency for at least some of the plants. Greenhouse experiments in which Rongelap soil was used as well as field trials at Rongelap have demonstrated that potassium fertilization decreases cesium-137 uptake by plants (Walker and Held, 1959). This observation is in agreement with the reported increase in cesium-137 uptake by plants in other soils as the potassium concentration in the soil is reduced by prolonged cropping (Nishita ef al., 1958 and 1959). It appears possible that if a surplus of potassium were present in the soils at Rongelap, potassium levels in Birgus muscle would be more uniform and cesium-137 levels lower. The simplest system studied at Rongelap with respect to cesium-137 and potassium was the soil. Cesium-137 and potassium in rain water which had percolated through undisturbed soils were found to be present in approximately