REASONS FOR INTERCHANGE OF CROPS. 275 



the quantity of phosphates extracted from the soil by the same 

 weights of wheat and tobacco must be as 97.7 to 16, and 

 when the difference is so great as this, the plants may suc- 

 ceed each other. 



4. Now if we examine what are called the ameliorating 

 crops, we shall find that they contain a very small quantity of 

 alkalies or of substances containing nitrogen, or of both. Thus 

 the leguminous plants contain only traces of salts, p. 266, and 

 hence they do not injure the crops of corn which are sowed 

 with or succeed them. The root crops require still less of 

 these alkalies and salts, and hence their ameliorating effects. 



5. If we observe the rotation which is carried on in na- 

 ture, for example, that pine trees succeed oaks, and oaks 

 pines, and examine their ashes, we shall find the reason of it. 

 *' One thousand parts of the dry leaves of the oak yield 55 

 parts of ashes of which 24 parts consist of alkalies soluble 

 in water," while the same quantity of pine leaves gives only 

 59 parts of ashes which contain 4.6 parts of soluble salts, 

 {De Saussure); and generally those trees whose leaves are 

 renewed annually, require from 6 to 10 times more alkalies 

 than the fir-tree or pine. 



6. It must be evident, without further examination, that 

 the causes of the failure of crops when cultivated successive- 

 ly on the same field, and the reasons for rotation, are to be 

 found in the kind and quantity of the substances, which each 

 species of plant extracts from the soil. Some agricultural 

 writers have held to the hypothesis, that each species of 

 plant requires different kinds of food, and when it has ex- 

 hausted its specijic food from the soil, another species will 

 flourish until its specific food is exhausted. We may learn 

 from the above examination what this specific food is. It is 

 the alkali* or salt which the plant requires for its develop- 

 ment. 



It should be remarked, however, that as one alkali may be 

 * See Liebig, p. 216. 



