=167. Determination of the temperature of condensation.= A detailed investigation was made by Rutherford and Soddy (_loc. cit._) of the temperatures at which condensation and volatilization commenced for the two emanations. The experimental arrangement of the first method is shown clearly in Fig. 59. A slow constant stream of gas, entering at _A_, was passed through a copper spiral _S_, over 3 metres in length, immersed in a bath of liquid ethylene. The copper spiral was made to act as its own thermometer by determining its electrical resistance. The resistance temperature curve was obtained by observation of the resistances at 0°, the boiling point of liquid ethylene −103·5°, the solidification point of ethylene −169° and in liquid air. The temperature of the liquid air was deduced from the tables given by Baly for the boiling point of liquid air for different percentages of oxygen. The resistance-temperature curve, for the particular spiral employed, was found to be nearly a straight line between 0° and −192°C., cutting the temperature axis if produced nearly at the absolute zero. The resistance of the spiral, deduced from readings on an accurately calibrated Weston millivoltmeter, with a constant current through the spiral, was thus very approximately proportional to the absolute temperature. The liquid ethylene was kept vigorously stirred by an electric motor, and was cooled to any desired temperature by surrounding the vessel with liquid air. The general method employed for the radium emanation was to pass a suitable amount of emanation, mixed with the gas to be used, from the gas holder _B_ into the spiral, cooled below the temperature of condensation. After the emanation was condensed in the spiral, a current of electrolytic hydrogen or oxygen was passed through the spiral. The temperature was allowed to rise gradually, and was noted at the instant when a deflection of the electrometer, due to the presence of emanation in the testing vessel _T_, was observed. The resistance, subject to a slight correction due to the time taken for the emanation to be carried into the testing vessel, gave the temperature at which some of the emanation commenced to volatilize. The ionization current in the testing vessel rose rapidly to a maximum value, showing that, for a small increase of temperature, the whole of the radium emanation was volatilized. The following table gives an illustration of the results obtained for a current of hydrogen of 1·38 cubic centimetres per second. Temperature Divisions per second of the electrometer −160° 0 −156° 0 −154°·3 1 −153°·8 21 −152°·5 24