Two environmental simulation chambers were used to study the behavior of a :3 *PuO ; heal source deposited in an arid, temperate environment. Large fragments from an impact-tested source were deposited in the first chamber, and fines from the source were placed in the second. Plutonium released into Xhe soil, the soil drainages, and the condensates from the dehumidifiers were monitored throughout the experiment. The bulk of the released plutonium, up to 0.04% of the source, was retained in the soil. Of the remainder, larger amounts of plutonium were released to the condensates (to the air) than to the soil percolates. Release rates to the condensates and percolates increased with time for both chambers. Condensate release rates averaged 23 nCi/day for the fines chamber and 90 nCi/day for the chamber with the large fragments. Percolate release rates averaged 224 nCi/rain for the fines chamber and 170 nCi/rain for the other chamber. Spalling of particles from the surface of the plutonium oxide to the air, caused by the action of rain on the hot source fragments, seemed to be a significant release mechanism for the large source fragments but not for the fines. A simulated cleanup operation, involving removal of the source and a small amount of soil, had little effect on the continuing release of plutonium to ihc soil percolates and to the dehumidifter condensates, except that the condensate release rates from the chamber that had contained the large source fragments dropped to a value equal to that of the other chamber. The large source fragments remained virtually unchanged after 8.5 yr of exposure, in contrast to one previous experiment in which fragments of similar size disintegrated into fines.