Atmometer for Ecological, Instrumentation. 81 



broad, so as to reduce the rate of change in the height of the 

 water cohmm as water is withdrawn. 



The instrument is installed as follows: Vials C and D are 

 supplied with clean mercury to a depth of 6 or 8 mil imeters 

 and tlieir stoppers tightly inserted. The stopper of the reservoir 

 G is raised, the bottle slipped to one side, and filled with dis- 

 tilled water. A rubber tube is attached to the free end of the 

 tube F and suction is applied, causing water to rise from the 

 reservoir, filling first vial C, then the porous cup, then vial D, 

 and lastly escaping into the rubber tube, which is then removed. 

 After slight adjustments of the mercury level, by raising or low- 

 ering one or both vials, the instrument is ready for operation. 

 A suitable support and a waterproof apron to protect all the 

 reservoir from rain may be readily arranged. To replace the 

 cup with a new one, the old cup is simply removed from its 

 stopper, allowing the contained water to escape, the new cup is 

 placed in position, and suction is applied to tube F as in installing. 



As water evaporates from the cup, mercury rises in tube B 

 and falls in tube E, until water from the reservoir passes freely 

 from E to C. The lost water is now supplied as usual from the 

 reservoir, the column of mercury in the lower end of tube B re- 

 maining at a height of a few millimeters. This height can be 

 varied by raising or lowering the vials. Water cannot enter the 

 cup from vial D and this vial remains filled. In time of rain, 

 evaporation from the cup ceases, it absorbs water from without, 

 the mercury rises in tube E and falls in tube B until water flows 

 freely from the cup through vial D to the exterior. No water 

 can enter the reservoir G from this direction, and the readings 

 obtained from it will give the actual evaporation. Rain-water, 

 (and dew also, if it plays a role here) is automatically shunted 

 to the outside. 



Various modifications of the instrument are of course possi- 

 ble, but the form here described seems most convenient when 

 ease of filling and changing the cups is considered. 



The error introduced by the volume change required for 

 the operation of the automatic mercury valves is so slight as to 

 be completely negligible in all climatic and habitat studies, and 

 this form of instrument is not apt to be used for finer studies, such 

 as those on transpiration. 



