2 74 THE PLANT WORLD. 



been established by experiment, yet it is well to restandardize 

 them at the end of a season's work. 



The instruments used in these experiments stood in tin 

 cylinders plunged in the soil to such a depth that the file 

 mark on the jar was level with the soil surface. 



The method above described for setting up the atmo- 

 meter allows rain to fall upon the cup but prevents its direct 

 entrance to the jar. As long as the outer surface of the cup 

 is covered with a water film, as is the case while it is raining, 

 the movement of water through the tube is reversed and 

 water actually flows into the jar, but at a very low rate, de- 

 termined by the height of the cup above the water level in 

 the jar. This introduces an error into the records, an error 

 depending, not on the am omit of precipitation, but on the 

 length of time during which the outer surface of the cup is 

 kept wet. Correction can be made for this feature if records 

 are taken of the duration of the periods of precipitation, rec- 

 ords which are also necessary for a proper study of the re- 

 lation of rainfall to soil moisture, and hence to plant growth. 

 Of course the error mentioned here can likewise be avoided 

 by placing a small screen over the cup, but such a screen 

 would alter the evaporation rate to some extent, even though 

 made of glass, and since the influx of heat from the sun 

 varies to such a great degree in difterent regions, such a 

 screen is undesirable. This error is so small that it is 

 neglected in these studies. 



Seeds for the experiment were sown in May and from 

 that time the drought conditions continually increased in se- 

 verity till July 6, when the season of summer rains set in. 

 From May 13 to July i the weekly average of the daily tem- 

 perature maximas rose from 87 degrees to 107 degrees 

 Fahrenheit, while the corresponding average of minima rose 

 from 45 degrees to 75 degrees. For the same period the 

 weekly averages of maxima and minima of relative humidity 

 decreased from 45 per cent (maximum) and 32 per cent 

 (minimum) to 31 per cent (maximum) and 17 per cent 



