AGRICULTURAL EXPERIMENTS 247 



lished by these experiments and that the results have passed 

 into the region of traditional farming practice. 



But to return to the nitrogen question. Lawes and Gilbert 

 attacked it in another fashion by repeating in the laboratory 

 certain experiments on the growth of plants in confined spaces 

 supplied with soil and fertilisers of known composition, so as to 

 be able to draw up a balance sheet of the nitrogen in crop and 

 soil as against the original nitrogen in seed and soil. Experi- 

 ments of this kind are subject to considerable error and no 

 conclusive results had been obtained by several previous in- 

 vestigators : but the Rothamsted experiments, which showed that 

 the nitrogen in the plant had been wholly derived from the 

 soil, were generally taken to demonstrate the fact that the living 

 plant cannot bring into combination the free nitrogen of the 

 atmosphere. As part of the same question, systematic measure- 

 ments were made of the ammonia brought down in the rain- 

 water, a large gauge measuring one-thousandth part of an acre 

 in area being constructed in order to obtain sufficient rain- 

 water for analysis. During forty years systematic analyses 

 have thus been made month by month, which show that, on 

 the average, the rainfall at Rothamsted brings down about 

 4 lb. of combined nitrogen per acre per annum, whereas the 

 ordinary crop will remove from 40 to 150 lb. per acre, thus 

 disproving Liebig's opinion that the rainfall provides a supply 

 of ammonia sufficient to maintain the crop without any external 

 assistance from nitrogenous manures. 



The leguminous crops provided certain facts inexplicable on 

 this theory of the source of the plant's nitrogen. Several of 

 the Rothamsted records showed that the growth of clover, 

 in addition to yielding a crop containing more than the 

 normal amount of nitrogen, also left the soil richer in nitrogen 

 than it was at the outset ; the clue to these facts was finally 

 supplied only in 1887, when Hellriegel and Wilfarth dis- 

 covered the existence of certain bacteria living " symbiotically " 

 upon the roots of clover and possessing the power of fixing 

 gaseous atmospheric nitrogen. Novel as was this conception, 

 it served to explain the anomalous results which had been 

 recorded at Rothamsted and the later work of Lawes and 

 Gilbert consisted largely in confirming and giving practical 

 shape to this discovery of Hellriegel and Willarth. Still later — 



