21 The pumping recovery tests, which involved pumping the borehole completely dry and then observing the rate of water level rise, conducted in the tremie material. These tests showed a very ra recovery in the water levels; in one case the water level rose in 5 minutes. Furthermore, water levels within the trenmie conc corresponded very closely to sea level and lagged only about 1/ behind outside tidal fluctuations. Thus, it appears that although the permeability of the soil mixture is quite low, the permeability of the tremie concrete i higher. Purthermore, in the tremie region there was relatively communication with the ocean, perhaps mainly along channels pr hy the oversize debris. The soil-cement mixture was a moist, dense material that cr in the hand and the tremie concrete, a dense, partially cement material. The whole of the crater contents, however, was rathe led impermeable to water except where there was channeling. In summary, the cores showed that there are zones of incomplktely cemented tremie concrete. This segregation of the concrete mos resulted because the tremie pipe was not always kept below the of the slurry, probably due to movements of the barge carrying injection equipment or failure to use a plug when each pumping was started. The soil-cement above the water level aiso did no achieve the concrete-like character that was anticipated, possi because of bacterial effects of organic material which prevent hardening of the concrete at the level of cement content used. theless, we believe that the keywall and concrete come are sati for all likely situations that will occur. 4.3 Radioactive Contents Samples taken from the cores and water samples frex the hol analyzed at the Lawrence Livermore Laboratory. Water samples a taken from two monitoring wells sunk outside the dome area, on lagoon side, in positions calculated to intersect water passing the crater and into the lagoon. The wells were fitted with unsc slotted polyvinylchloride pipe for use in future monitoring. The results of these analyses have been given by Robison and Noshkin (1981). To summarize the data here, mean values have be calculated, omitting samples taken in the fall-~back zone. Resul given in pci/g and the range gives the high and low values for t of samples. Strontium and Pu were analyzed using wet chemistry and the others, using gamma ray spectrometry. Radionuclide 239 T2s¢py 7. Am **Sr 13768 Mean Concen* cation Range 18.6 46 - 1.6 2.8 6.3 - 0.20 20.6 §2 ~ 5.5 8.7 27 - 0.24 rface