I* 



temperature are distributed over so many successive daily 

 periods that their effects on the accuracy of this re- 

 lation between reservoir-loss and plant-intake are neg- 

 ligable. There are, of course, progressive changes 

 in water content of the soil as the season changes 

 from cold to warm, or vio p versa , but these do not af- 

 fect the accuracy of the daily relation between loss and 

 intake. In the experiments here reported it was as- 

 sumed that the twenty-four-hour water-loss from the re- 

 servoir was equal to the intake by the plant over the 

 same period. It is possible to test the reaction of a 

 particular pot and soil to varying temperature and thus 

 to obtain and apply a correction, but this is an extreme 

 of accuracy which is not justified by the other condi- 

 tions of the experiments. 2he reasons for this will be 

 obvious when the experiments are described • 



In most cases no precautions were taken to re- 

 move oxygen from the water supplied through the auto- 

 irrigators. However, in those experiments in which it 

 was important to come as close as possible to excluding 

 all traces of oxygen, the water-reservoir of the auto- 

 irrigators we're provided with water from which the free 

 oxygen had been removed by boiling. In this case, the 

 boiled water was kept in a reservoir provided with an 

 atmosphere of nitrogen, and the water-reservoir of 

 each individual auto-irrogator was also provided with 

 a nitrogen atmosphere. Usual devices were employed 



