In order to follow the rapid changes in the heating effect of radium, after removal of the emanation, Rutherford and Barnes (_loc. cit._) used a pair of differential platinum thermometers. Each thermometer consisted of 35 cms. of fine platinum wire, wound carefully on the inside of a thin glass tube 5 mms. in diameter, forming a coil 3 cms. long. The glass tube containing the radium and also the tube containing the emanation were selected to slide easily into the interior of the coils, the wire thus being in direct contact with the glass envelope containing the source of heat. The change in resistance of the platinum thermometers, when the radium or emanation tube was transferred from one coil to the other, was readily measured. [Illustration: Fig. 100.] The heating effect of the radium in radio-active equilibrium was first accurately determined. The radium tube was heated to drive off the emanation, which was rapidly condensed in a small glass tube 3 cms. long and 3 mms. internal diameter. After allowing a short time for temperature conditions to become steady, the heating effect of the radium tube was measured. The results are shown in Fig. 100. An observation could not be taken until about 12 minutes after the removal of the emanation, and the heating effect was then found to have fallen to about 55 per cent. of the maximum value. It steadily diminished with the time, finally reaching a minimum value of 25 per cent. several hours later. It is not possible in experiments of this character to separate the heating effect of the emanation from that supplied by radium A. Since A is half transformed in three minutes, its heating effect will have largely disappeared after 10 minutes, and the decrease is then mainly due to changes in radium B and C. The variation with time of the heating effect of the active deposit is still more clearly brought out by an examination of the rise of the heating effect when the emanation is introduced into a small tube, and of the decrease of the heating effect after the emanation is removed. The curve of rise is shown in the upper curve of Fig. 101. 40 minutes after the introduction of the emanation, the heating effect had risen to 75 per cent. of the maximum value which was reached after an interval of about 3 hours. [Illustration: Fig. 101.]