Is Flow Verification Necessary?

Safeguards test statistics are used in an attempt to detect diversion of special nuclear material. Under assumptions concerning possible manipulation (falsification) of safeguards accounting data, the effects on the statistics due to diversion and data manipulation are described algebraically. A comprehensive set of statistics that is capable of detecting any diversion of ma terial is defined in terms of the algebraic properties of the effects. When the assumptions exclude collusion between persons in two material balance areas, then three sets of accounting statistics are shown to be comprehensive. Two of the sets contain widely known accountancy sta tistics. One of those sets does not require physical flow verification - comparisons of operator and inspector data for receipts and shipments. The third set contains a single statistic which does not require physical flow verification. In addition to not requiring technically difficult and expensive flow verification, this single statistic has several advantages over other comprehensive sets of statistics. When the as sumptions do not exclude collusion, then none of the three sets of statistics is comprehensive. This algebraic approach to safeguards accountancy is expected to be useful in the further study of collusion and other diversion scenario problems.