326 The Physiology of Plants BOOKIII 



views were strengthened by the discovery that, when he 

 was experimenting with a difference of electrical potential 

 not greater than that which frequently exists between 

 strata of the atmosphere not far from the ground, filter 

 paper which had been moistened with water had its nitrogen- 

 content increased fourfold. On substituting a solution of 

 dextrin for the water the amount of nitrogen fixed increased 

 to sixteen times the original quantity. 



These observations stood by themselves till 1885, when 

 Berthelot became alive to the presence and activity of 

 a new agency, supplementing the purely physical pheno- 

 mena of his first experiments. He found during that year 

 that certain clay soils fixed atmospheric nitrogen under the 

 influence of the presence of certain microbes or bacteria, 

 which, however, he did not determine, though he stated 

 definitely that he held the process to be rather due to them 

 than to the soil in which they were living. He published 

 further results in 1892. 



This was an important discovery, for though it pointed 

 to the atmosphere as a great source of nitrogenous supply 

 it indicated no less clearly that the utilization of it is in 

 no sense a function of the chlorophyll apparatus, and that 

 therefore we are face to face with some kind of chemo- 

 synthetic process ; it thus opened up a number of problems 

 as to the relation of the two. Moreover, as the fixation 

 appeared to be due to the protoplasmic activity of the 

 microbe, it emphasized the likelihood that we must dis- 

 criminate between the function of the colouring matter and 

 that of the living plastid in the process of photosynthesis. 

 Further, it showed how the green plant may be ultimately 

 supplied with nitrogen from the air, though, as Boussingault 

 and other observers had shown, there is no evidence of 

 any direct absorption. 



Naturally so striking a pronouncement as this of Berthelot 

 was followed by a long series of experiments which were 



