6 Prof. Magnus on the Propagation of Heat in Gases. 



was placed in a glass cylinder P Q, 235 millims. wide and 

 400 millims. in height. This stood in a second similar cylinder 

 XY, so that there was a space of 30 millims. on every side, 

 which was filled with water. In order that the internal cylinder 

 P Q might not be raised by this water, flat leaden weights were 

 placed on its base, which loaded it so that it rested on the cork 

 supports U U. This cylinder was closed at the top by a hollow 

 metallic cylindrical cover, E E, 75 millims. deep, in which water 

 was poured. In this cover there was a cylindrical aperture d d' s s 1 , 

 55 millims. in diameter. The vessel C which received the bailing 

 water was fitted in this aperture by means of a slit cork; it was 

 prevented from falling by a couple of metal slides, s s 1 , fitted on 

 the under surface of the cover. This arrangement served at the 

 same time to fasten the whole apparatus ABC. 



The heat which the vessel radiated laterally, heated the water 

 in the cover E E. In order to keep it at an invariable tem- 

 perature, cold water continually flowed from a high reservoir 

 through the tube rr, while the heated water escaped by a siphon 

 hh. To observe the temperature on the inner cylinder there 

 were several thermometers, one of which, k k, was placed in a 

 horizontal position right under the cover EE;a second, m, was 

 suspended in the middle of the space, and a third, /, near the 

 bottom. During the experiment all these thermometers were 

 kept at the same temperature, 15° C. For this it was necessary 

 to have the room in which the experiments were performed at 

 about this temperature. 



The bulb of the inner thermometer fg was protected by a 

 screen o o from direct radiation from above. At first I used a 

 cork screen, but afterwards one of silvered copper foil. Accord- 

 ing to the kind of gas contained in A B, the thermometer reached 

 its highest point and kept it unchanged, in from twenty to forty 

 minutes from the time at which hot water was poured into the 

 vessel C and steam passed into it. In the same kind of gas, 

 the density being constant, the maximum was always reached 

 in the same time; and provided that the thermometer fg, with 

 its screen, always remained in a fixed position and distance from 

 the vessel C, the temperature did not vary more than 0°1 to 0°2C. 

 Under these conditions also the gradual increase of the tempe- 

 rature, up to the maximum, took place in such a manner that 

 after the same time the thermometer always indicated the same 

 temperature. This concordance furnished a proof of the accuracy 

 of the method. 



Before turning to a few of the conclusions from the numbers 

 obtained, it will be convenient to enter upon the circumstances 

 which influence the maxima of temperature. 



The heat proceeding from the lower surface of the vessel C is 



