W. od. Bair June 18, 1979 Page 6 or estimating Sr-90 in bone from that in diet. concept of correlation is not involved. The statistical I suggest the authors do not use the word “correlate” since bone data are not availab 18. p. 8 Ys t paragraph This paragraph needs to be rewritten. Why is Bennett's model thought to best reflect Sr-90 concentrations in bone at Enewetak? from? What alternative models are there to choose Bennett's model was developed and fit to Sr-90 concentrations in diet and in mineral bone as determined from autopsy samples. However, this paragraph reads as if output from the model calculations somehow involve autopsies. Also, Bennett's model will use estimated not actual diet information for Enewetak application. 19. p. 8 Last 4 lines The authors should explain more fully why Martin and Bloom's model, which lumps "the upper two long term compartments into one,” is acceptable for Enewetak application. Why is the difference between the ICRP and Martin and Bloom's model not significant? 20. p. 8, Line 4 What information is available on AMAD. When was an AMAD of 0.5 used? 21. Pp. 9 Lines 11-14 Why were 15% and 85% chosen as the short and Jong term intake parameters for the two compartment Cs-137 model? Cs is probably the single most important radioisotope on Enewetak. Yet, the authors give only 2 or 3 lines to describing model parameters! 22. 23. p. 10, Line 5 What is meant by the phrase “Engebi is the limiting living from botton pattern of those evaluated"? p. a I've emphasized above the description of the diet survey results and the methodology of the diet survey are completel) inadequate. The sample questionnaire displayed in Figure 2 leads me to question the validity of the survey results. For example, what was the definition of "normal" and "famine" conditions applied to the Enewetak people? Under