One purpose of the soil samples is to determine the Pu/Am ratio in order to caleulate Pu concentrations from 24lAm concentrations. The total concentrations will include all Pu isotopes for which Eberline determines values. Preliminary data indicates that, for most islands, the set of ratios is distributed symmetrically and unimodally, with small variance; the mean of the distribution is therefore the desired ratio value. Histogram plots, goodness-of-fit tests, or other analyses will be used to verify the shape of the distribution and estimate the mean. On a few islands, the ratio distribution has a large variance, or is a mixture of two or more distributions with different means, If possible, the island will be divided into subsections so that each contains ratios from a pure distribution. Statistical analyses will be performed to verify the appropriateness of the subdivision, and additional samples requested as necessary to assure accurate results. If this proves impossible, soil samples would, as a last resort, have to be taken at every survey location. Documentation concerning the ratios used, the areas each ratio applies to, and justification for each will appear in the daily log and the results notebook. The chosen ratios will be used to calculate Pu concentrations, on which the covariance structure will be refitted if necessary. Another objective of the soil sampling is to confirm the calibration factor on the in situ detector. The average 241Am from soil samples should roughly equal the in situ value; since the actual area of measurement of the two methods is much different, exact equality is unlikely. If, however, the two values are totally inconsistent, EG&G and Eberline will be informed immediately so that the soil samples and in situ data can be checked. It is imperative that such discrepancies be resolved before any additional sampling is done, IV. Procedures for Other Data Bases For the health physics data base, Eberline will produce data stored on tape as two strings, which will be written logically to disk, one sample per two-record file (EICDB2 on disk; tape file 26). File name, disk label and catalog procedures are similar to the in situ data base. Details, along with data formats and sample retrievals, are in the health physics data base documentation. Source documents of data collected by the FRST are maintained by the JTG staff, and will be used to input that data by hand to a separate FRST data base*, The data, two strings per sample, will be stored logically on one-record files, one sample per file. Field data from contaminatedislands and environmental data from clean islands will be stored in the same format, but on separate disks. Because of the increased probability of error due to hand input, a printed copy of the input data will be made, checked against the source document, and retained permanently. Details on file names, disk labels, catalogs, and sample retrievals are in the FRST data base program documentation. Vv. Other Analyses, Documentation and Maintenance Procedure Statistical analysis may be required on other types of data (e.g., water consumption patterns); the type of analysis appropriate to the situation is a matter of judgment for the Statistician. The plotter should prove an effective tool for presenting data and results, and for producing special format reports. Complete, accurate documentation will be maintained continually. For example, permanent alterations in a program will be stored on the tape and disk copies and the program listing and documentation and the tape and disk catalogs updated. New programs in the repertory will be stored, listed and documented, and placed in the program documentation notebook. Originals or copies of results of covariance fits, estimates, or other analysis will be stored in the results notebook, along with explanatory documentation as required. The daily log will contain notes on work aecompushed, programs written or revised, problems encountered, approaches and suggestions for the other statistician. *Responsibility for entry of FRST data was transferred to a military base in the U.S. in the fall of 1977; thus, DRI had no further contact with the FRST data after October 1977. A-3-3