of dark Heat-rays by Gases and Vapours. 11 



holes like a sieve; so that / is continually exposed to a current 

 of water, which flows over the edge of h and runs off by the 

 tube s. It need hardly be remarked that the temperature of 

 the water was observed from time to time. In the course of 

 months a slight change occurred ; but during the time 

 occupied by an experiment the temperature was absolutely 

 constant. 



After the temperature of the whole has become uniform, 

 the little plate a is suddenly to be increased in temperature by 

 a certain amount. We sought to accomplish this by suddenly 

 directing a current of steam against a. In the arrangement 

 of the heating-apparatus it was necessary above every thing 

 to consider that the thin copper plate a in a series of experi- 

 ments must always be equally heated. But since, in consequence 

 of the close neighbourhood of the large quantity of cooling- 

 water, the loss of heat by conduction was considerable, it was 

 only possible to secure equality in the radiation towards the 

 thermopile (which was to be suddenly set in action) by always 

 directing upon a equal quantities of vapour with equal velocities, 

 and also in the same direction and at the same distance from 

 a. Hence it was necessary that equal quantities of vapour 

 should be generated in equal times. Hence, when ebullition 

 commences, there must always be nearly the same quantity 

 of water in the apparatus M employed for the generation 

 of steam, and the flame should always be of the same size. 

 The first object was attained by employing the arrangements 

 for keeping M filled with water to the same level represented 

 in the plate. The flask K, about five sixths full of water and 

 inverted over the funnel T, has two tubes, passing air-tight 

 through an india-rubber stopper, which dip below into the 

 water in the funnel, and which terminate at different levels 

 inside the flask, one just inside the stopper, the other reaching 

 into the air in the upper part of the flask. This arrangement, 

 as is well known, secures a constant level in the water in T. 



Each time, on beginning to heat the vessel M, the tube was 

 closed at o by means of a clamp, which was opened again at 

 the end of the experiment. In order to have a constant source 

 of heat under M, a manometer, m, containing petroleum, was 

 used to measure the pressure of the gas supplying the burner 

 z, the current of gas being regulated by means of the screw- 

 clamp y. Another manometer, d, indicated accidental excess 

 of pressure in M. The production of vapour was allowed to 

 become so rapid that the steam issued briskly from the 

 end of the tube v. This tube began and ended in glass tubes 

 which were connected by a piece of caoutchouc tubing ; and 

 the whole was wrapped round with linen to prevent premature 



