In this apparatus, which was first constructed in 1905, heat supplied by radiation is directly compensated by the peltier absorption of heat in a thermo-junction through which a measured electric current is passed. In the simplest form of the instrument, radiation admitted through a measured aperture 2 mm. diameter falls on a small copper disc 3 mm. diameter by 0.5 mm. thick, to which two thermo-junctions are attached, forming a peltier cross. One couple is connected to a sensitive galvanometer for indicating changes of temperature. The other is connected to a battery and rheostat in series with a milliammeter or potentiometer for measuring the current required to reduce the deflection of the galvanometer to zero. If A is the area of the aperture in square centimetres, H the intensity of the radiation in watts per square centimetre, a the absorption coefficient of the surface of the disc, P the peltier coefficient in volts, C the balancing current in amperes, and R the effective resistance of the couple, the equation giving the value of the radiation in absolute measure is aAH = PC - C2R. The absolute value of P is the product of the absolute temperature by the thermoelectric power. The value of R, in the small correction term for the joule effect, is readily determined by observing the neutral current, C0 = P/R, for which the joule effect balances the peltier effect. In practice two similar discs with similar connections are mounted side by side in a thick copper box, and are balanced against each other in order to avoid changes of zero due to exposure to sunshine, or rapid variations of temperature. The advantages of the disc radio-balance are that it is very simple to construct and easy to reproduce without material variation in the reduction constants. It is very suitable for measurements of solar radiation, or strong sources, but is insufficiently sensitive for weak sources; and the absorption coefficient a must be determined by comparison with a standard. In the cup radio-balance the radiation is received in a copper cup 3 mm. diameter by 10 mm. deep, so that the absorption coefficient is practically equal to unity. Greater sensitiveness is secured by employing a pile of several couples, insulated from the cup, in place of the single balancing couple. External disturbances are eliminated by employing a pair of cups, similarly mounted but oppositely connected, enclosed in a thick copper cylinder. The joule effect, represented by the C2R term in the equation, is automatically eliminated by passing the same current in series through the opposing peltier junctions soldered to the bottoms of the cups. The cup exposed to radiation is cooled, and the cup screened from radiation is heated by the peltier effect, while both are equally heated by the joule effect. A complete observation involves reversing the current and switching over the radiation screen, in order to eliminate any difference of sensitiveness of the two piles. By observing the neutral current, each cup can be used separately, as with the disc balance; but the disc balance cannot be used with the peltier couples connected in opposition, unless the balancing couples are insulated from the discs. The cup radio-balance is sensitive to less than a tenth of a microwatt, and is very suitable for measuring the heat evolved by small quantities of radio-active substances. It was applied to radium at Prof. Strutt's suggestion, and Prof. Rutherford very kindly supplied samples of emanation, and determined the value of the radium sample employed by comparison with his own standards. The second sample of emanation had only just come to hand, and the absolute values had not been finally reduced at the time the paper was read; but it appeared from the preliminary reductions that the heat evolution of radium in terms of Prof. Rutherford's standards was somewhat greater than that given by previous observers. The instruments exhibited at the meeting included (1) a disc radio-balance equatorially mounted for solar radiation; (2) a laboratory pattern, with water or steam jackets for determining the temperature coefficients and for investigating the theory of the instrument; (3) a cup radio-balance, with which the heat production of radium and its emanation were demonstrated. The latter intrument is capable of measuring the heat evolved from 1 milligram of radium to about one part in 1,000.