41 



From the beginning I attempted to avoid this danger. 

 In the dark room the gas-main was turned off permanently. 

 The thermostat was heated electrically. Moreover, the dark 

 room was ventilated several times a day, by means of 

 the electric fan, the fresh air being obtained from the 

 spacious rooms and corridors of the laboratory, where 

 no gas was burned either, as the whole laboratory is 

 lighted electrically. The air passing through the thermostat 

 Itself came directly from the roof of the laboratory. 



In this way I thought sufficient précautions had been 

 taken against any impurity of the atmosphère, in which 

 I had to make my experiments, but expérience has taught 

 me not to be too readily satisfied in this respect. 



The thermostat was the Trojan horse, which brought 

 in the bad vapeurs unnoticed. As a warning to others 

 I will describe my expériences on this point somewhat 

 more fully, especially because this source of error has 

 in the aggregate detained me at least for one month. 



In the last days of December 1909 the thermostat 

 arrived; it was constructed of copper, the inside was 

 covered with tin and the outside painted and at that time 

 I began to work in the thermostat in the dark room. 



Before leaving on January P' 1910 for a ten days' 

 holiday I made some preliminary experiments at 20° C, 

 25° C. and 30° C, after a preliminary warming of 1 hour. 



Kesuming my experiments on January 15"' I could not 

 succeed in getting the same presentation-time at 80° C, 

 which I had found on December 31''. I thought of every 

 possible disturbing influence which could make itself felt 

 inside the thermostat, I thought of phototropism, of 

 hydrotropism, of aërotropism, of thermotropism, but not 

 one of the expérimental variations I tried could throw 

 light on the cause of my failure. At last I was quite 

 certain that the cause was not any polarity of my apparatus. 



