"Records of the Geological Survey of India. 
262 
[vOL. Xllli 
Formation of kankar. 
close relation between the two. Tlie relatire proportion of common salt to 
sodium sulphate was found by Medlicott to vary from 4 to 24 per cent. 
17. The re-deposit of carbonate of lime gives rise to tho.se nodules known as 
kankar. It tabes place at the upper margin of the im¬ 
permeable subsoil. They are not formed by the lime 
depositing round a nucleus and pushing the other elements of the soil aside. A 
jjortion of rather porous soil, consisting of a mixture of lime, sand, and clay, is 
infiltrated with water retained in it by an impermeable bottom. The carbonate 
of lime is deposited tbronghont this porous mass, and cements its particles toge¬ 
ther till it becomes of a stony hardness. Deposit no doubt also takes place along 
the outer surface, as each former minute crystal deposited acts as a nucleus for 
further depo.sit. The formation is often seen in an incomplete state, nodules of 
soil having become only partially hardened. The process is essentially one of 
segregation fi’om the soil itself. Such nodular fonnations, which are very com¬ 
mon with other minerals, as iron oxide, silica, &c., are an example of the simjilest 
kind of metamor23hosis going on in rocks and soils. It is not necessarily con¬ 
nected with efflorescences on the surface. The essential condition of its existence 
i.s the presence of carbonate of lime, or its ready production by ordinary decom¬ 
position in the soil. In soils and sub-soils which supply little lime there may be 
effloroscencos without formation of kankar, as in those consisting of clay and 
silicious sand. On the other band, in marly soils, in which there may be little 
production of alkaline salt, kankar may form without any efflorescence. The 
analysis of kankar very w’oll illustrates their mode of formation. They show 
from 20 to 50 per cent, of carbonate of lime, the rest consisting of the mixture of 
clay and sand of -which the soil is composed. 
18. To estimate practic.ally the amoirnt of injurious roh in any soil, it should 
Estimation of rcli in be washed with boiled distilled water and the solution 
soils and water. eraporatedj then burned to expel organic matter and 
finally weighed. In the case of the w’aters of rivers, canals, and w’olls, they 
should bo evaj)oratod, ignited, ro-carbonated, and w'cighed. The easily soluble 
salts should then be washed out with a little distilled water and the residue 
■weighed. The portion undissolved consists of lime and magnesian carbonates 
and some sulphate of lime with small amounts of silica, &c. The difference 
between the two weights is the amount of salt capable of efflorescing. If one 
have a record of the analy.sis of any water, a rough approximation is got by 
deducting from the total solids the volatile matter (almost all organic), also the 
removable hardness consisting of carbonate of lime. In addition two grains per 
gallon of carbonate of lime should bo further deducted, as in boiling (in order 
to remove carbonate of lime), two grains per gallon still remain dissolv’ed. A still 
further deduction would require to be made for silica, iron, &o., but these are in 
small amount. I mention these methods of approximate estimation because they 
are readily applied and are useful for all 23ractical pur230scs. 
19. In considering the conditions that lead to accumulation of salts on the 
Causes of aocnmula- surface or in the tindorground water, it is to be borne in 
tioii of salts. mind that soils exposed to moisture, air, and he.at are 
continually generating them, und that in some in which the felspathie elements 
