104 



Development and Activities of Roots of Crop Plants. 



pare with those at the beginning of the experiment. Great care was exer- 

 cised to secure thoroughly representative composite samples for nitrate 

 analysis. A small amount of toluene was added at once to the sample in the 

 air-tight jars to stop bacterial action. 



Nitrates were determined by the reduction method, using Devarda's alloy 

 and the method proposed by Whiting, Schoonover, and Richmond (1920). 

 The results of these analyses, together with the data on root extent and water- 

 loss, are summarized in table 37. Final analyses for nitrates were made only 

 in those fertilized soil layers into which the roots extended. 



Table 37. — Loss of water and nitrates'^ from November 4, 1920, to March 12, 1921. 



1 The nitrate-content is given in parts per million of NO3. This includes not only the NOi 

 of the NaNOs added, but all NO3 present in the soil when analyzed at the beginning and end 

 of the experiment respectively. 



^ This container was taken down on December 21. 



A study of table 37 shows that the roots developed only poorly; in only 

 3 of the 12 containers did they reach depths greater than 2 feet. Moreover, 

 they were scarcely more than half as abundant as in the earlier field experi- 

 ment (p. 94). These findings lead the writers to seriously doubt the value 

 of many greenhouse experiments with crop plants when the results are applied 

 to field conditions. Not infrequently the indoor environment is made still less 



