[en] Practical thresholds for water leak-tightness of small leaks were determined by experimentation. Measurements for small leak samples were taken of air leakage rates and water leakage rates for identical leak samples in order to identify parameters that influence water leak-tightness threshold. Four types of leaks were evaluated: a fine wire inserted in an O-ring seal, a glass capillary tube, a stainless steel orifice, and a scratched valve stem on a 1 inch UF₆ cylinder valve. Experimental results demonstrated that the key parameter for water leak-tightness is the opening size of the leak hole. The maximum allowable hole size to achieve water leak-tightness ranged from 10 to 20 μm in diameter in this study. Experimental results with 1 inch UF₆ cylinder valve samples demonstrated that the acceptance criteria for preshipment leakage test, 1x10^-3 ref-cm³.s^-1, as prescribed in ANSI N14.5 is an appropriate value from the point of view of water leak-tightness for enriched UF₆ packages. The mechanism of water leak-tightness is plugging by tiny particles existing in water. The water used in experiments in this study contained far fewer particles than in water assumed to be encountered under accident conditions of transport. Therefore, the water leak-tightness threshold determined in this study is a conservative value in a practical evaluation. (author)