CHAP, iv The Ash of Plants 361 



1856-60, and consisted in cultivating plants in insoluble 

 artificial soils, to which he added the particular materials 

 whose fate he was endeavouring to trace. By its means 

 he considered he had determined that phosphorus, sulphur, 

 potassium, calcium, magnesium; iron, silicon, and manganese 

 are essential to the wellbeing of all green plants, and 

 especially of the cereal grasses. He was doubtful about 

 chlorine, and denied that sodium is of any utility at all. 



This method was modified and improved by Hellriegel 

 in 1883 ; he introduced the use of sand that had been 

 heated to redness and then boiled in sulphuric acid. The 

 salts under observation were added subsequently. 



The method of water culture, which had originally been 

 used by,_Woodward as long ago as 1699, was revived in 

 1860 by Sachs, and was used in the same year by Knop. 

 This method, which consists in growing plants in a nutritive 

 solution containing known proportions of various salts, 

 offers many advantages over the others and has been the 

 basis of most of the investigations that have been made 

 since it was agair^ introduced. The results obtained by 

 its employment have established certain facts as to the 

 ash constituents which call for notice, though most of the 

 knowledge that has been obtained has been of the most 

 empirical nature. They have led to little more than specu- 

 lation, and hardly any accurate knowledge as to the specific 

 functions of any of the ash constituents. 



The method of water culture enabled Nobbe to deter- 

 mine in 1862 that for one plant at least, the buckwheat, 

 chlorine is essential for full development. He found that 

 with an appropriate solution he was able to cultivate this 

 plant with complete success, indeed, he found it to flourish 

 more luxuriantly than in ordinary soil. Wolff found in 

 1868 that oats behaved similarly. 



In 1866 Birner and Lucanus found that the elements 

 deemed essential by Salm-Horstmar should not have in- 



