The several water intrusion scenario studies in the recent literature are all quite similar and may be easily understood if used to estimate the total number of eventual cancers per unit of energy generated, including their sensitivity to input parameters. However, these studies are grossly overpessimistic in several aspects of the problem, especially in using leach rate data from highly unrealistic experimental situations, and in ignoring geochemical considerations in both leaching and in transport. It is concluded that it is reasonable to expect removal and transport for an atom of buried waste to be similar to that for an atom of average rock. Under that assumption, the leach rate can be estimated from the chemical compositions of rock and of groundwater, coupled with the water flow through aquifers. The result (excluding 238U) is 0.0008 eventual cancer/GW(electric)-yr. This treatment would be invalidated if the waste were released through fractures in the rock induced by the emplacement operations or by heat. If such fractures cannot be discounted, total reliance must be on leach resistance.