92 . WIELOPOLSKI, ADAMS, AND HEOTIS exposure of some Marshallese to radioactive fallout in 1954 (10). Nevertheless, the program encompasses many aspects of health care not related to radiation. Specimen collection. Nonfasting blood samples were collected at the time of medical examination in New York, Rongelap, or Majuro. The venipuncturesite was cleansed with 70% isopropyl alcohol (Tomac Prep Packet, Amer. Hosp. Supply, Edison, N.J.). Multisample Vacutainer blood collection tubes Model B-D 6453, containing disodium EDTA, and Model B-D 6512 (Becton, Dickinson and Co., Rutherford, N.J.) were used for red cells and serum, respectively. Packed red cells were obtained after washing the anticoagulated blood three times with normal saline (0.9% sodium chloride injection, Travenol, Deerfield, Ill.). Red cells and sera were frozen in polyethylene chemical containers (Bel-Art, Pequan- nour, N.J.) for transport and storage. The sametype of vial was used for storage of atoll water and coconut liquid collected for the evaluation of environmental sources of Br in drinking fluids. Specimen analysis. In preparation for analysis 300-ul samples of thawed serum or red cells were layered onto Parafilm over an area of 1 cm?. The specimens were then covered to avoid dust contamination andair-dried. Sample processing was alternated among the groups of specimens to minimize systematic bias in preparation or analysis. For EDXRFanalysis the sample was irradiated with soft X rays emitted by a 1099Cd radioactive source. This technique creates vacancies in the inner atomic shells of the elements. The subsequent emission of characteristic X rays by the target atoms is monitored by a Si(Li) detector. The signals are processed by standard electronics and displayed in the form of a pulse—height distribution. This method enables quantitative and qualitative evaluation of the sample composi- tion. In the present study, red cells were analyzed for Cl, K, Fe, Zn, Br, and Rb, and the sera were analyzed for Cl, Zn, and Br. Other elements were below the detection limits of the EDXRF system employed. Calibration. The EDXRF system was calibrated for Br with the addition method in which serum and red cell samples were spiked with graded amounts of Br (5 to 15 g/ml) in the form of NaBrOQ;. It was determined that the mean value of 3296 counts of Br corresponds to a concentration of 1.3 pg Br/ml of packed red cells (the calibration line is ppm = 4.02 x 10-4 XRF —4.83 x 10-4 where XRF is the Br peak intensity) and a mean value of 23,971 counts corresponded to 6.4 pg Br/ml of serum (calibration line is ppm = 2.62.10-+ XRF + 8.33.10-2). These values were used to calculate absolute Br concentrations in the blood samples. Statistical analysis. For analysis of normalized values of all the detected ele- ments, a standard one way analysis of variance (ANOVA) wasused. The test was applied to test the equality of group means, and statistics were computed to test the equality of means between each pair of groups. The nonparametric Wilcoxon rank sum test and Spearman’s rank correlation were used for analysis of group differences in serum and red cell Br levels. RESULTS Possible contamination of containers was investigated. Vacutainers are known to be contaminated with inconsequential amounts of Br (<0.006 and <0.05 wg per 5012840