THE CHEMICAL COMPOSITION OF LAKE WATERS 153 
magnesium chloride. But if the waters subjected to concentration 
are alkahne, the result will be very different. As concentration 
proceeds the preponderating sodium carbonate tends to throw not only 
calcium but also magnesium out of solution, and the final liquors will 
consist almost entirely of sodium salts, viz. carbonate, sulphate, and 
chloride. In certain rare instances boric acid originating from 
volcanic vents has found its way into lakes. Its presence, in solution 
seems to be confined to highly saline alkaline lakes, and this may be 
due to the fact that where calcium and magnesium are present in 
appreciable quantity, boric acid would tend to be eliminated as 
insoluble borates ; whereas in alkaline concentrates it would persist in 
solution as borax (sodium pyroborate). 
During the world's history many lakes must have dried up 
completely after accumulating a large store of salts. In moderately 
humid climates this cannot have happened often, but when it did 
happen, an inverse process of re-solution must have gradually set in. 
Thus the saline residues would lose first magnesium and then sodium 
salts, whilst calcium sulphate and carbonate might well survive into 
recent geological periods. Rock-salt deposits generally, and especially 
the sodio-magnesio-potassic deposits of the North German Plain, are 
monuments of bygone lakes of sea- water, cut off from the ocean ; pro- 
bably, however, these are instances not of desiccation to the last 
drop, but of copious deposition of salts followed by withdrawal of the 
mother-liquors. Far less resistance is offered to the formation and 
survival of saline residues in arid regions ; many such, of very variable 
composition, are known to exist, some of them being exploited com- 
mercially, especially in the Nile Valley, Central Asia, and the United 
States. Since arid regions, as we have seen above, are apt to produce 
alkaline waters, these deposits consist as a rule largely or mainly of 
sodium carbonate, occasionally with a considerable proportion of 
borax. 
Of a very different class of solute, which is never absent in lake 
waters, viz. the dissolved gases, there is but little to be said. Whilst 
this department of hydrology has received a great deal of attention 
from oceanographers, experimental data as to the gases dissolved in 
lakes are, so far, scanty and isolated ; and it is to be admitted that 
the subject bristles with physical and chemical complications, and 
presents no small experimental difficulties. Pure water in contact 
with air takes up oxygen, nitrogen, and carbon dioxide up to definite 
limits of saturation. The amount of each gas taken up is directly 
proportional to its partial pressure, decreases, though not in a simple 
relation, with increasing temperature, and lastly depends on a solubility 
constant which varies somewhat widely from gas to gas. As an effect 
of their respective solubilities, oxygen and nitrogen go into solution 
