ATOMIC HYDROGEN 103 



worth while avoiding them. For example, if you have in lamps a vacuum 

 as good as you know how to produce, but suspect that the lamps would be 

 better if you had a vaciuim, say, lOO times as good, it may be the best 

 policy, instead of attempting to devise methods of improving this vacuum, 

 to spoil the vacuum deliberately in known ways, and you may then find 

 that no improvement in vacuum is needed or, just how much better the 

 vacuum needs to be. 



During these first few years, while I was thus having such a good 

 time satisfying my curiosity and publishing scientific papers on chemical 

 reactions at low pressures, I frequently wondered whether it was fair that 

 I should spend my whole time in an industrial organization on such purely 

 scientific work, for I confess I didn't see what applications could be made 

 of it, nor did I even have any applications in mind. Several times I talked 

 the matter over with Doctor Whitney, saying that I could not tell where 

 this work was going to lead us. He replied that it was not necessary, as far 

 as he was concerned, that it should lead anvwhere. He would like to see me 

 continue working along any fundamental lines that would give us more 

 information in regard to the phenomena taking place in incandescent lamps, 

 and that I should feel myself perfectly free to go ahead on any such lines 

 that seemed of interest to me. For nearly three years I worked in this way 

 with several assistants before any real application was made of any of my 

 work. Tn adopting this broad-minded attitude Doctor Whitney showed 

 liimself to be a real pioneer in the new type of modern industrial research. 



For my study of the efifect of gases, I had to devise new types of 

 vacuum apparatus. I needed particularly to be able to analyze the small 

 quantities of gas that existed in the tungsten lamp. With some of this 

 special apparatus I was able to make a practically complete quantitative 

 analysis of an amount of gas which would occupy about i cu. mm. at 

 atmospheric pressure. In tjjis sample of gas we could determine the per- 

 centages of oxygen, hydrogen, nitrogen, carbon dioxide, carbon monoxide, 

 and the inert gases. 



In regard to the fate of the different gases which I introduced into the 

 lamp bulb, I found that no two gases acted alike. Oxygen attacked the 

 filament and formed tungstic oxide, WO3. That seemed simple enough, 

 but the kinetics of the reaction presented many features of considerable 

 scientific interest. 



In studying the effect of hydrogen very peculiar phenomena were ob- 

 served. A limited amount of hydrogen disappeared and became adsorbed 

 on the bulb, where it remained in a chemically active form, which was 

 capable of reacting with oxygen at room temperature even long after the 

 tungsten filament had been allowed to cool. This suggested hydrogen atoms 

 and seemed to confirm some conclusions that I had already drawn from 



