15 to about 400 eps as the counter descended through the fall-back zone and rose again to about 3400 cps at the true bottom of the crater. The high count at the surface of the fall-back zone is probably due to the fact that small particles, which absorb more activity per unit of weight than large particles, fell back more slowly. The high permeability of the coral rock ensures that the radio- active material within the fall-back zone has been continuously leached by sea water since 1958. Nevertheless,- substantial quantities of radioactive material were present beneath the apparent bottom of the crater before any of the soil and debris from the islands was placed into it. It is also possible that a part of Cactus Crater was formed out of a man-made extension of the island on the lagoon side of the reef (Defense Nuclear Agency 1981, p. 409); at least there is no appreciable beachrock present on the lagoon side of the crater. 3.9 Pillingof the Crater The contaminated soil was transported by barge to Runit Island, where it was mixed with cement and attapuigite to form a mixture designed for use in the tremie method of underwater concrete placement. Using this method, water is added to the cement-soil mixture to form a slurry that is pumped through a pipe to the underwater location; the end of the pipe is kept below the surface of the ejected slurry to prevent segregation of the cement and soil. The crater was filled to the low-tide water level using the tremie method. The key-wall then was sunk to a depth of 1 foot where the beachrock was solid and to a depth of 8 feet where the beachrock was fractured or absent. The key-wall apparently was placed by deposition through water that inevitably entered the forms because of the high permeability of the formations on which the key-wall was placed.” Above the water level, a common soil=cement placement method was used in which a layer of contaminated soil was spread and bags of cement were placed at designated intervals and punctured. The cement was blended inte the soil with a dise and the layer was compacted. Using this procedure a dome-shaped mound was formed over the crater. Radicactive debris (i.e., metallic debris, contaminated concrete, and other large pieces of material) too large to pass through the tremie pipe later was placed in an area, (called the “donut hole”) reserved ‘for it in the center of the structure and was "choked" in place with slurry. Before the filling of the crater was completed, construction of the concrete cap or dome was started. It consists of 358 panels in ll rings, and the panels vary in size from 20 by 2r feet at the outside te 6 by 7.5 feet near the center. The panels were made in place in forms and rested on polyethylene sheet. The design thickness of the panels was 18 inches, but the actual thicknesses ranged from 12.5 to 24 inches, with a mean of 17.3 inches (Ristvet 1980). The outer ring "Segregation, however, was observed in the core samples only at te en hod er cs the bottom of the key-wall.