Duane — Heat Generated by Radio-active Substances. 261 



rium is reached, when the heat conducted away from the 

 calorimeter equals that given it per second by the source, and 

 observe the maximum displacement of the bubble of air. This 

 method works well provided the instrument is not arranged 

 for very great sensitiveness. 



If, however, the apparatus is very sensitive it is better to 

 take the velocity of the bubble as a measure of the heat 

 generated per second. Although the instrument is well pro- 

 tected against thermal disturbances from the outside, yet the 

 bubble does not stay in the same place. The zero of the 

 instrument is not fixed. Nevertheless, if the apparatus has 

 remained undisturbed for a long time, and the temperature 

 throughout has become as nearly equalized as possible, the 

 natural drift of the bubble is slow and regular, and the change 

 in its velocity due to the heat from the source, when it is 

 lowered into the calorimeter, can be measured with consider- 

 able precision. 



A third method is to compensate the effect of the heat 

 generated in the tube D by generating a known quantity of 

 heat in the corresponding tube J)' (figure 1). 



The best method, however, is to compensate the heat effect 

 by absorbing the heat in the tube D itself as fast as it is 

 generated. This can be done by means of a current of 

 electricity flowing across the junction of two metals. Peltier 

 discovered that if the current passes in one direction heat is 

 generated, and if in the opposite direction heat is absorbed at 

 the junction. 



In my earlier experiments I inserted a thermo-couple P of 

 iron and nickel wires into the tube D, and I determined the 

 current that absorbed the heat as fast as it was generated, by 

 varying the strength of the current until the velocity of the 

 bubble was the same as its natural drift. In the later experi- 

 ments I have replaced the simple thermo-couple by a metal 

 tube. The walls of the tube are l mm thick, and its external 

 diameter is just enough less than the diameter of the tube D 

 to allow of its being inserted easily into the latter. The length 

 of the metal tube is about 4 cm , so that the entire tube lies 

 inside the calorimeter. Half of the tube is of iron and the 

 other half of nickel, the two surfaces between the two metals 

 being vertical and parallel to the axis of the tube. An iron 

 wire is soldered to the outer edge of the iron half of the tube 

 and a nickel wire to that of the nickel half, so that a current 

 of electricity descending by the iron wire into the iron half of 

 the tube can pass across the joints into the nickel half and 

 ascend by the nickel wire. With' this arrangement, when a 

 source of heat is lowered into the middle of the iron-nickel 

 tube, it is surrounded by a good conductor of heat, and the 

 distribution and compensation of the heat takes place easily 



