292 THEORIES OF FERTILISER ACTION [chap. 



solution of considerable concentration' possessing a 

 proportionally increased solvent power, and it is to this 

 supersaturated solution that may be attributed the 

 highly localised attack of the roots upon the soil 

 particles. An experiment by Kossowitsch illustrates 

 the part played by the roots in attacking the insoluble 

 materials in the soil : two pots of sand were prepared, 

 each mixed with the same quantity of calcium 

 phosphate in the form of ground rock phosphate, a third 

 pot contained sand only. In this latter and in one of 

 the pots containing the calcium phosphate, seeds of 

 mustard, peas, and flax were sown. The growing plants 

 were then furnished with a slow continuous supply of 

 water containing appropriate amounts of nitrates, 

 potash, and other nutrient salts except phosphates. 

 Before, however, this nutrient solution reached the pot 

 containing the sand only, it was made to percolate 

 through the second pot containing sand and calcium 

 phosphate, but it was applied directly to the pot con- 

 taining calcium phosphate. In the pot containing 

 calcium phosphate, the growth was much greater than 

 in the other pot, where the nutrient solution only con- 

 tained what phosphoric acid it could dissolve in its 

 passage over the calcium phosphate in the pot in which 

 nothing was growing, although this solution was 

 continually renewed. The only factor determining the 

 supply of phosphoric acid and the consequent difference 

 in growth was the solvent action of the roots when 

 they were actually in contact with the calcium 

 phosphate, and this solvent action, as has already been 

 shown, may most probably be attributed to the carbon 

 dioxide they excreted. 



Following up their conclusions that the soil water 

 possesses an approximately constant composition under 

 all circumstances and always contains more of the 



