130 Tj^ansactions of the Boyal Society of South Africa. 
able conditions, dissolve in the rain-water and drain out of the soil, but they 
accumulate and act injuriously where circumstances are unfavourable to 
their removal by natural drainage. During a rainy season the salts are 
carried to varying depths according to the penetrating capacity of the 
rain-water. The surface soil is thus left free from injurious salts. When 
dry weather sets in, however, the soil water rises to the surface as oil 
rises in a wick by capillary attraction, carrying the noxious salts with it 
from below\ The water evaporates from the surface soil, leaving behind 
the salts as a white incrustation. There is, hence, a constant downward 
passage of these injurious alkaline salts during rain, and an equally 
constant upward movement during dry weather. 
It will be easily understood that irrigation will tend to increase the 
accumulation of alkaline salts in a soil, seeing that the irrigation water 
will probably penetrate to greater depths than the rain, and carry to the 
surface larger quantities of salts. The danger is naturally augmented 
when the water used for irrigating is itself alkaline. 
Hence, when adaptability of any tract of country for irrigation has to 
be pronounced upon, two chemical problems have to be decided : firstly, 
does the soil contain any constituent which may render it brack or other- 
wise unproductive ? and secondly, assuming a satisfactory answer to the 
first point, is the land sufficiently provided with the necessary components 
of the food of plants to make farming profitable ? 
When the amount of plant food in a soil is the only point at issue, 
one needs not to probe to a greater depth than the first foot or 18 inches, 
but it is obvious, from what has been stated above, that in order to ascer- 
tain not only whether a soil is at present unproductive through brackness, 
but also whether it is liable to become so when irrigated, it is not enough 
to take merely a sample of soil from the surface and analyse that. It is 
essential to take samples at regular intervals below the surface to a depth 
at least equal to that w^hich the irrigation water may possibly reach. 
There should be no omission that may leave a loophole for mistake. 
For this reason the method employed is to take samples representing 
successive sections vertically downwards from the surface to a depth of 
from 4 to 6 feet. Generally 1-foot sections are taken, but in certain cases 
10-inch, 8-inch, or even 3-inch sections have been collected as occasion 
seemed to demand. Sometimes it has been advisable to sample solid 
blocks of soil, for instance, in the Thebus investigation, where the 
following method was used : — 
A hole, several feet in diameter, was excavated to a depth of 6 feet, a 
solid pillar of earth being left untouched in the centre. From this central 
pillar six samples were taken in each case, every sample representing the 
average of 1-foot depth of soil. It will be readily seen that such a method 
of sampling involves more labour than scraping a couple of handfuls of 
