31 



In the ceiling an electric fan is placed above a case of 

 black cloth, so that the room could be aired quite suffi- 

 ciently in a short time. 



As the dark room is sltuated in the centre of the 

 building, sudden changes in the température out of doors 

 were felt only very slowly. Moreover a small steam radiator 

 could be used, while on the other hand the fan together 

 with spraying in the room and the space between the 

 doors with water made it possible to keep the température 

 of the room some degrees below the température of the 

 building if desired. 



During winter the room was kept damp by watering 

 once or twice a day, later on even this was unnecessary 

 as the water of the thermostat and of the water-current, 

 which worked the air-pump, were sufi6cient to keep the 

 room damp. At each experiment I noted the humidity of 

 the room measured in degrees of a hygrometer, which 

 had previously been checked at the points 0° and 100°. 

 Of course this way of measurlng was rather a rough one, 

 but it was sufflcient for my purpose. I tried to keep the 

 humidity at 70° or higher, seldom it was below 60°. 



§ 7. Sources of error. 



As it was my purpose in this investigation to obtain 

 results quantitatively as trustworthy as possible, I 

 bestowed particular care on the tracing of ail sources of 

 error and on their élimination. The first four months of 

 the investigation were taken up almost entirely with 

 this and the first 300 experiments (of which a few only 

 are mentioned in this paragraph) were used almost 

 exclusively to get a clearer insight into thèse difficulties. 



The results of this preliminary investigation form the 

 contents of this section. For the sake of convenience 



