THE WORK OF ROOTS 159 



ALKALI SOILS. The tests for alkali are of importance in arid and 

 semi-arid regions. Whenever the ra nfall is small the salts are not 

 leached out of the soil so fast as they are formed and they accumulate. 

 As the rainwater evaporates they are left on the surface ( see experiment 

 on page 127), where they often form a whitish deposit which disappears 

 during a rain but reappears as soon as the soil begins to dry out. These 

 salts may be in the form of "black alkali," or sodium carbonate (so called 

 because it combines with the humus to form a black mass) ; such "black 

 alkali" spots are very conspicuous. This is the worst form of alkali, for 

 it corrodes the plant just at the surface of the soil and kills it. By adding 

 gypsum (sulphate of lime) the "black alkali" is changed to "white 

 alkali," Glauber's salt (sodium sulphate) and carbonate of lime. Glau- 

 ber's salt, together with common salt (sodium chloride), magnesium salts, 

 etc., are called "white alkali": they are much less injurious than "black 

 alkali." Quantities of alkali salts are often found at varying distances 

 (down to several feet) below the surface of the soil, where they form 

 compact masses known as alkali hardpan ; its presence may be ascertained 

 (like that of any hardpan) by sounding the soil with a sharp-pointed 

 steel rod to the depth of four or five feet at least or by digging holes. 

 When we begin to irrigate such land the alkali of the hardpan often 

 rises and the alkali spots on the surface gradually become larger 

 and larger. The experiment already described on page 127 shows 

 clearly how this takes place ; in this case the salt in the bottom of the 

 tumbler represents the alkali hardpan. This rise does not take place if 

 the irrigation is properly done ; only excessive application of water 

 brings it about. This suggests that we might reverse the operation by 

 flooding the land with water and carrying it off by means of drains placed 

 underground. This would correspond to placing the tumbler of sand used 

 in the experiment under the faucet and allowing the water to run out of 

 a crack in the bottom ; it is very evident that the salt would soon dis- 

 appear. Large areas of alkali lands have been reclaimed in this way, 

 which furnishes the only practical method of reclamation. To wash out 

 the salts completely would be a disadvantage, since they consist, to a 

 considerable extent, of substances needed by the plant. The water of 

 moderately alkaline lakes can be used for reclaiming alkali soils, if enough 



