288 THEORIES OF FERTILISER ACTION [chap. 



given by the authors of the theory, of the cold water 

 extracts from a number of soils show great variations in 

 their concentration in nitrates, phosphoric acid, and 

 potash; nor is any evidence forthcoming that such 

 concentrations are not immediately raised by the 

 addition of fertilisers. Indeed, when the Rothamsted 

 soils, with their long-continued differences in fertiliser 

 treatment, are extracted with water charged with 

 carbon dioxide — the nearest laboratory equivalent to 

 the actual soil water — the amount of phosphoric acid 

 going into solution is closely proportional to the 

 previous fertiliser supply, and this proportionality is 

 maintained if the extraction is repeated with fresh 

 solvent, as must be the case in the soil. In the field it 

 is not merely the initial concentration of the soil water 

 in plant food which determines the supply of nutriment 

 to the crop ; it is also the capacity of the soil to keep 

 renewing the solution as the plant withdraws from it 

 the essential elements. 



In one essential respect again the conditions pre- 

 vailing in the soil are very different from those of the 

 laboratory. In the soil all reactions are extremely 

 localised, since they take place in the thin film of water 

 normally surrounding the soil particles, in which 

 movement of the dissolved matter takes place very 

 slowly, mainly by diffusion. Of the extreme slow- 

 ness of the diffusion of soluble salts in the soil the 

 Rothamsted experiments afford some good examples. 

 For instance, on the grass plots only an imaginary line 

 divides the plots receiving different fertilisers ; the 

 manure is sown right up to the edge of the plot, a 

 screen being placed along the edge to prevent any 

 being thrown across the boundary, then immediately 

 on the other side of the boundary the different treat- 

 ment begins. In two cases plots receiving very large 



