EVAPORATION RATES FOR SHORT INTERVALS 139 



a very fine capillary glass tube, the other was a small porous 

 clay cup attached by a rubber stopper to the free end of the 

 supply tube. Thus all the water passing from the reservoir to 

 the atmometer cup had to pass through the capillary or, in the 

 other case, through the porous walls of the little cup. Cylindri- 

 cal atmometer cups of the usual form were employed and the 

 manometer was made of glass barometer tubing having a bore 

 of about a millimeter. These instruments operated very satis- 

 factorily at first, but the resistance to water movement always 

 increased with the lapse of time, apparently because of some 

 sort of clogging of the capillary or of the small porous cup, 

 so that readings on the manometer scale soon became unreliable. 

 During the winter of 1915-16, in this Laboratory, the writers 

 experimented further with this kind of apparatus, attempting 

 to devise a form of resistance that would not readily alter, but 

 the attempt was unsuccessful and the whole matter has been 

 laid aside. 



If the instrument with a pressure gauge can not as yet be 

 produced, it behooves us to employ the nearest approach thereto 

 that is readily feasible, and such an approach has been carried 

 out by merely employing a device by which the ordinary pair of 

 readings may be made with a very short intervening time period. 

 The essentials of this device are as follows. The atmometer cup 

 is mounted so as to draw water from a finely graduated burette — 

 a straight pipette graduated to hundredths of a cubic centimeter 

 is satisfactory — and this burette is also connected with the 

 regular water-reservoir, but a cock allows water movement 

 through this connection to be out off entirely. When the cock 

 is open the water level in the burette and that in the reservoir 

 are the same (or but slightly different because of capillarity), 

 and the instrument draws from the combined supply in the 

 usual way, readings for long periods being made by filling 

 through the small burette, from a larger burette or graduated 

 cylinder. When the cock is closed the instrument draws 

 water solely from the small burette, by means of which readings 

 may be made in terms of hundredths of a cubic centimeter. 

 Of course the cock must be reopened before the small burette 



