NUCLEAR-DEBRIS FORMATION 25 ues were obtained by first constructing a mass chain-yield curve from literature values judged to be the most reliable and then distributing the chain yield among the chain members according to the method of Coryell.4 The reliability of the values obtained in this way is difficult to estimate and probably varies from one fission case to another. The chain yields are known quite well for thermal-neutron fission of **U but are probably subject to some revision in all other cases, It is con- ceivable that such revisions might sometimes involve sizable factors, The distribution of the chain yields among the membersof the chain, charge distribution, is even more uncertain. Again, the values for thermal fission of *°U are most reliable since they are supported by considerable experimental evidence. The experimental values for 3513 thermal fission are, in fact, the principal basis of the charge- distribution method. For the other fission cases, data to support the calculated distributions are sparse. An important gap in the independent-yield information is the lack of knowledge concerning the partition of independent yields among isomeric states—usually the ground state and an excited state. When no experimental data are available, the computer program assumesthat the independent yield of the nuclide is divided evenly among the iso- mers. This is not expected to be correct, and recent reports’’’”® of ex- perimental studies along these lines indicate that it is not. Description of Decay Chains The nuclides comprising a decay chain (mass chain) must be specified, including isomeric states, and the parent— daughter relations indicated. These are mostly well known, of course, but some uncertainties involving isomeric states still exist. Half-lives and Branching Fractions The values for half-lives and branching fractions now in use are listed in Report USNRDL-TM-1937.* An uncomfortable percentage of these (26% of the half-lives and 18% of the branching fractions) are estimates. Most of the half-life estimates are taken from Bolles and Ballou.” In addition, about 15% of the half-lives are not even estimated but are simply assigned an arbitrary value of 1 msec. These all pertain to early chain members of very low inde- pendent yields and very short half-lives. The 26% that were estimated involve values that range from a few seconds to a few minutes. Expe- rience has shown that these estimates may be in error by as much as a factor of 100. The branching-fraction estimates were made mostly by assigning equal probability to two or more branching paths. No way is known to approximate these values more accurately. *G. R. Crocker, R. C. Scheidt, and M, A. Connors, Radionuclide Input Data for Fission-product Abundance Computations, U. S. Naval Radiological Defense Laboratory Technical Memorandum (USNRDL-TM) No. 137, San Francisco, Aug. 6, 1963. USNRDL-TM’s are not generally available for distribution.