W. J. Bair/R. O. Gilbert July 9, 1979 Page 2 There is not sufficient data to predict coses from aquatic foods; i.e., clams, crabs, or fish. Livermore dismisses the doses as insignificant, but under certain circumstances, the Marshallese diet contains large quantities of fish. Other researchers have mentioned fish, etc. as pathways deserving of attention. I think more dai: or analysis of aquatic foods should be presented to back up any assumptions made. No data are presented at all for radionuclide concentrations on the southern islands of the atoll. These data are important, since for the first 8 years the diet of the northern islands is supplemented by crops grown on the southern islands, and some scenarios use southern~island-grown crops. I have been assuming that all concentrations are negligible in comparison with the northern islands; i.e., essentially zero. For the ingestion pathways, I used exclusively our modifications of the ICRP-2 methodology. Therefore, our doses have some inherent differences. I have checked their models against ours and can reconcile almost all differences easily. Our models include a total-body contribution: from ?°Sr. Joe Soldat tells me that of 0.32, our 29Sr doses compare well (between *10-25%). This is pretty good the ICRP-2 model I use The Bennett model used hard mineral bone. If mineral bone and apply probably overestimates the 29Sr total-body contribution. by LLL is only for bone marrow. Our bone model is for I use our model for predicting the 2°Sr concentration in the bone/marrow dose ratio of Spiers (1972) and UNSCEAR considering totally different uptake models are used. My doses from !37Cs + D are uniformly higher than those presented by LLL. This is because I use the ICRP-2 model and they use the ICRP-10 double-exponential model. This results in their long-term dose commitments being 85% of ours. If this 0.85 is factored in, I am within +10% of their answer for total body. My 127Cs bone results are still slightly higher, since, for my model, bone has a 140-day biological halflife as opposed to a 115-day biological halriife for total body used by ICRP-10. The only radionuclide I am having trouble reconciling is 2*Am. My ingestion doses are about one order of magnitude lower than LLL's, even if I take their increased gut-to-blood transfer coefficient into account. A preliminary check of inhalation dose from **!Am does not show this problem. While this may reflect that I am only approximating the air concentration, I am going to have to more closely examine our actinide model and parameters. I have only a few minor complaints about the rest of the dose presentation. There is no concentration ratio data for arrowroot plants. These plants become a major fraction of the diet during famine conditions.