92 University Geological Survey of Kansas. 
Hypothetical Combination.—As it has been the custom among 
water analysts to report the constituents as being present in 
definite combination as salts, the ‘‘hypothetical combination ’’ 
has been stated in another column, so as to make the analysis 
more intelligible to the ordinary reader and to the physician 
who is not familiar with ions or radicals. This is especially 
desirable because we can thus compare the analysis of those 
waters with the analyses of other chemists. The problem in 
regard to what combination should be made is a complex one, 
for certain combinations that would take place in a dilute solu- 
tion do not take place in a concentrated one. If the solutions of 
two salts are mixed, and an insoluble salt is precipitated by this 
mixture, then the reaction is a simple one. But how shall we 
know what takes place when, for instance, solutions of mag- 
nesium sulfate and sodium chlorid are mixed? No precipitate 
is formed. Does the solution contain each salt as originally 
present, or does a metathetical reaction take place, forming so- 
dium sulfate and magnesium chlorid? 
So many factors must be considered in the discussion of this 
question that we must decline to answer it positively. We 
must know the effect of temperature, of concentration, and of 
other salts in solution, as well as of the relative quantities of 
each salt present, before we can form a judicial opinion as to 
what combinations are actually present. If the solutions are 
dilute no change takes place, and the substances remain as ions. 
If the solutions are concentrated, all possible salts as well as 
all the ions will be present. (For a further discussion of this 
problem in the light of the most modern theories, see chapter V.) 
Following the example of most chemists, the results are ex- 
pressed in terms of grams per liter. This seems to be prefer- 
able in the case of mineral waters where the amount of total 
solids is large, rather than expressing the results in terms of 
parts per hundred thousand or parts per million, according to 
the custom in reporting the sanitary analysis of waters. We as- 
sume, of course, that grams per liter represents parts per 
thousand by weight, but this is only true in the case of waters 
having a very low specific gravity, and to make it absolutely true 
