43 



expérimental objects caused by the stay in tlie room 

 never occurred afterwards. 



Of course tlie atmosphère in the durk room was not 

 absolutely pure; absolute purity is not to be attained in 

 a laboratory; but I felt fully satisfied, if the' air was as 

 pure as the conditions of my work would allow, and if 

 the remaining impurities had a constant influence without 

 making a comparison between the results impossible. 



As has been mentioned above, the atmosphère in the 

 dark room was thought to be pure, when the présentation - 

 time at 20° C. amounted to 4'20". Later on this changed : 

 after xMay 18'" the presentation-time at 20° C. decreased 

 to 3' (cf. tables VI and VU at 20° C. in Chapter II). 

 I feel pretty sure this has to be attributed to the fact, that 

 the building was not being heated any more after that 

 date, and was ventilated much better than in winter. 

 Perhaps also the fact was of some influence that from 

 February to May ^ the thermostat was heated day after 

 day ; after the 18"" of May on the other hand it was cooled 

 in order to work at low températures. As my starting- 

 point was the presentation-time of 4:'20" at 20° C, I hâve 

 multiplied the presentation-time of ail further observations 

 (since May IS"") by ^Va, by which raeans thèse results 

 became comparable to those obtained before May 18'*'. 



That the presentation-times at ail températures were 

 changed in the same way by this altération in the external 

 conditions was. proved by comparing not only experiments 

 at 20° C. before and after May 18"", but also experiments 

 at 0° C. The first détermination at 0° C. was made from 

 February 26'" to 29'", the second from June 17'" to 20'". 

 In the first case the presentation-time was 72', in the 

 second one 48' and "/s X 48' = 69'20". The agreement 

 between thèse values is very striking, when we take into 

 account the fact, that the first détermination was made 



