152 PHENOMENA, ATOMS, AND MOLECULES 



the rate at which the hydrogen molecules reach the surface may be cal- 

 culated. If, then, the heat loss be determined by experiment, the energy 

 carried away by each molecule can be calculated. If, by heating the filament 

 to very high temperatures, a condition could be reached in which every 

 hydrogen molecule which strikes the filament becomes dissociated, then 

 those experiments would lead to a direct determination of the heat of 

 formation (heat absorbed by dissociation) of hydrogen molecules. In the 

 absence of definite evidence that the dissociation is complete, the method 

 gives at least a lower limit for the heat of formation. 



EXPERIMENTS ON THE HEAT LOSSES FROM TUNGSTEN 



WIRES IN HYDROGEN 



I. Higher Pressures. — In the experiments upon which the former cal- 

 culations of the dissociation were based, the measurements of heat loss 

 were made from short pieces of tungsten wire mounted vertically in a large 

 glass tube open at the lower end, through which a fairly rapid stream of 

 hydrogen passed. The temperature in most cases was determined from the 

 resistance. The relation between resistance and temperature was, how- 

 ever, found by measurement of the intrinsic brilliancy of pieces of the same 

 wire mounted in exhausted bulbs. 



In the present experiments, it was desired to employ a series of different 

 pressures of hydrogen around the filament, and furthermore, to avoid 

 injury to the wire by impurities in the hydrogen. 



The method adopted was therefore to mount single loop filaments of 

 pure tungsten wire in large heater lamp bulbs (cylindrical bulbs about 

 25 cm. long and 7 cm. diameter) which were filled with pure, dry hydrogen 

 at various pressures. These were then set up on the photometer bench and 

 a series of simultaneous measurements of candle power, current and voltage 

 were made. The color of the light emitted was also accurately matched 

 against that from a standard lamp viewed through a special blue glass.^^ 



Before filling the bulbs with hydrogen, they were exhausted to o.i 

 micron pressure for an hour while heated to 360°, and the filaments were 

 heated to a high temperature to drive ofif gases. The hydrogen was pre- 

 pared electrolytically and was freed from oxygen and water vapor with 

 extreme care. 



Thirty lamps were made up for these experiments. Fifteen of these 

 were filled with hydrogen at the following pressures: i, 10, 25, 50, 100, 

 200 and 750 mm. Six were exhausted to a good vacuum in order to de- 

 termine the amount of energy radiated. The remainder were filled with 



^^ This method has proved extremely accurate and serviceable as a means of com- 

 paring the temperatures of tungsten filaments. It is described in some detail by Lang- 

 muir and Orange, Proc. Amer. Inst. Elect. Eng., 32, 1895 (1913). 



