140 EARL S. JOHNSTON AND BURTON E. LIVINGSTON 



is emptied, when it will refill to the level of the water surface 

 in the reservoir, or slightly above; it is not essential that the 

 burette be filled to the zero mark of its scale. The diagram of 

 figure 1 shows one way of erecting the apparatus. 



Such an instrument may be placed in a room or greenhouse, 

 or may be carried from place to place. It may be allowed to 

 operate indefinitely so long as considerable distilled water is 

 always present in the reservoir bottle and so long as the cock is 

 open. Readings may be taken in the ordinary way at hourly, 

 daily or weekly intervals, from which the average evaporating 

 power of the air for the period in question may be readily ob- 

 tained. When it is desired to make a determination for any 

 short period the cock is closed, the time required for the loss of 

 one or more hundredths of a cubic centimeter is determined by 

 direct observation and the hourly rate is obtained there rom. 

 Thus, if the time required for a loss of a hundredth of a cubic 

 centimeter were 30 seconds it would mean that the average rate 

 of water loss during that short period had been 1.2 cc. per hour. 

 In general, 36 divided by the number of seconds required for the 

 loss of a hundredth of a cubic centimeter of water gives the 

 hourly rate of water loss. If the temperature of the porous cup 

 changes considerably during the time required for the loss of a 

 unit of water a more or less pronounced error will of course be 

 introduced, on account of thermometer action; since a small 

 change in the temperature of the relatively large water-mass in 

 the cup produces a comparatively large alteration in the level 

 of the water meniscus in the burette. On account of the com- 

 paratively short time periods required this temperature change 

 is not generally apt to be considerable, however. 



