apparent hindrance. The problem before us is to determine whether such free molecules have any power whatever to stop the waves of heat, and if so, whether different molecules possess this power in different degrees.
The source of waves which I shall choose for these experiments is a plate of copper, against the back of which a steady sheet of flame is permitted to play. On emerging from the copper, the waves, in the first instance, pass through a space devoid of air, and then enter a hollow glass cylinder, 3 feet long and 3 inches wide. The two ends of this cylinder are stopped by two plates of rocksalt, this being the only solid substance which offers a scarcely sensible obstacle to the passage of the calorific waves. After passing through the tube, the radiant heat fails upon the anterior face of a thermo-electric pile, where it is instantly applied to the generation of an electric current. This current conducted round a magnetic needle deflects it, and the magnitude of the deflection is a measure of the heat falling upon the pile. This famous instrument, and not an ordinary thermometer, is what we shall use in these enquiries, but we shall use it in a somewhat novel way. As long as the two opposite faces of the thermo-electric pile are kept at the same temperature, no matter how high that may be, there is no current generated. The current is a con-
