When the relation between temperature and quantity of energy radiated is known, any instrument which will indicate the amount of the radiations it receives may be used to measure temperatures. The ray, for example, may be focused on a thermal junction, which will be heated in proportion to the amount of energy incident upon it, and when connected to a millivoltmeter will cause deflections proportional to the energy it receives. A thin strip of metal might be used in place of a junction, and by measuring its resistance the heating effect of the radiations, and hence the amount thereof, may be deduced. A third method would be to focus the rays on to a compound strip of two metals, which by altering in shape could be made to furnish a clue to the quantity of energy received by it. In theory, it is only necessary to allow the radiations to fall on the working part of any instrument for measuring low temperatures, when the rise in temperature produced may be taken as proportional to the energy received, and the thermal condition of the radiating body deduced from the fourth-power law. In practice, however, it is desirable that the receiving thermometer should be small in size; of low thermal capacity, so as to respond rapidly; and capable of giving a sensitive indication—hence an ordinary mercury thermometer would be unsuitable for this purpose. A thermopile, placed at a fixed distance, would fail owing to the cold junctions gradually warming up by conduction through the pile. The part receiving the radiations should be coated with lamp-black, so that practically all the waves impinging upon it, whether luminous or non-luminous, may be absorbed, and the energy they represent utilised in producing a rise in temperature.