Model tests of tanks are carried out by the experimental design method. Flow rate, nozzle height, and nozzle diameter of inflow gas are selected as elementary factors which affect the gas purging efficiency. The results of variance analysis are as follows. 1) The case that difference of densities of inflow gas and initial gas in tank is small. Flow rate is, only, considered as significant factor at 5% level of significance. The relation to purging efficiency is shown in fig. 10. 2) The case that the density of inflow gas is larger than the initial gas density in tank. Nozzle height is, only, considered as significant factor at 5 % level of signficance. The relation to efficiency is shown in fig. 11. 3) The case that the density of inflow gas is smaller than initial gas density in tank. Nozzle height and nozzel diameter are considered as significant factores at 5% level of significance. The relations to puring efficiency are shown in fig. 12.