1 68 SOIL CONDITIONS AND PLANT GROWTH 



the investigator to explain with some degree of completeness the ob- 

 served water relationships of the soil when sufficient is known about 

 the water supply, and also to account for many of the peculiarities ob- 

 served in cultivation. It enables him to say, as far as can be said on 

 our present knowledge, whether any observed defects are due to defects 

 in the soil or its situation, or to the system of management that has 

 been adopted. As it cannot be interpreted fully without a knowledge 

 of the amounts of organic matter and calcium carbonate present these 

 two quantities must be determined in every sample. 



We have seen that there is a close correlation between the potash, 

 the alumina and the clay. For purposes of a survey it seems super- 

 fluous to determine these two bases in every sample taken. The iron 

 oxide shows a general but by no means so close a correlation with the 

 others; but no connection could be traced between iron oxide and 

 fertility in the soils examined by the author, the iron oxide being 

 almost always less than 5 per cent, in amount. Nor did it appear 

 that the ratio of lime to magnesia in these soils was significant. The 

 nitrogen is closely correlated with the organic matter, i.e. the loss on 

 ignition. The total phosphoric acid does not show very great varia- 

 tions in the different soils, but the available phosphoric acid, like the 

 available potash, varies greatly with the management of the soil. Thus 

 the figures obtained by chemical analysis, apart from the loss on ignition 

 and the calcium carbonate, fall into two groups : the nitrogen, potash, 

 and alumina, which are so closely correlated with quantities already 

 determined in the mechanical analysis that their separate determination 

 is almost superfluous ; and the iron oxide, magnesia, lime, etc., which 

 do not give sufficiently useful indications to be worth determining in 

 every case. Since chemical analysis fails to characterise the soil with 

 sufficient completeness Hall and Russell recommend that for purposes 

 of a survey a large number of soils should be submitted to mechanical 

 analysis, including the determination of organic matter and of calcium 

 carbonate, and then a carefully chosen representative set should be 

 analysed chemically. They agree with Whitney that mechanical 

 analysis should form the basis of the survey, because it alone takes 

 account of those physical functions the regulation of the water supply 

 and therefore of the temperature, of the air supply, ease of cultivation, 

 etc. that play so large a part in determining the value of a soil. 



But, on the other hand, mechanical analysis is restricted in its appli- 

 cation and breaks down altogether on chalk soils, acid humus or peat 

 soils, and neutral humus soils, while it gives useful indications only on 

 the mineral soils, i.e. sands, loams and clays. Agricultural soils belong 



