4 
COLORADO EXPERIMENT STATION 
in the form of a solution, but it is a question whether this solution is the 
same in strength and character as the water in the drain. I think that I 
have clearly shown in Bulletin 65 of the Colorado Experiment Station, page 
34, that the water soluble portion of the soil is different from the salts held 
in solution in the ground water. Further in Bulletin 72, page 28 et seq. that 
the ground waters are different from the drain waters, and the action of 
these waters on the tile would vary somewhat according to the source of the 
solut’on which penetrated the mass of the tile. 
The drain waters obtained, say at a depth of four feet, contain as a rule 
a smaller amount of salts in solution than the ground water. The salts 
present are the same but their relative quantities vary. The usual salts 
present in the ground water's are, sodic sulfate, calcic sulfate, magnesic sul¬ 
fate, sodic carbonate, and sodic, in some instances also magnesic chlorid. 
In the drain waters we have the same salts but their relative quantities are 
different, the most notable feature being the relatively large amount of 
sodic carbonate. 
The substances claimed to be the most active in effecting the disinte¬ 
gration of cement are the sulfates. In the case of sea water the explana¬ 
tion offered is that the lime of the cement decomposes the magnesic salts 
present in the water" whereby the lime goes into solution and the magnesia 
is deposited in its stead, causing a bulging and disintegration of the work 
owing to its greater bulk. The principal magnesian salt in sea water is the 
chlorid, the deleterious action of which on cement is not established, but it 
has been shown that magnesic sulfate acts energetically.* The most active 
agent in decomposing concrete is the sulfuric acid of the sulfates carried in 
solution. Again solutions of gypsum have been shown to act detrimentally 
on concrete by forming with the tri-calcic aluminate a sulfo-aluminate. 
The water acting on these tiles whether it be ground water or drain 
water is very bad. One of the two small samples sent to me contained 
1252.6 grains per imperial gallon, of which 493.5 grains were sulfuric acid 
(SO3) with an extremely large proportion of magnesia. The other con¬ 
tained 542.3 grains per imperial gallon, with 42.31 grains of magnesic oxid 
or 7.8 per cent, of the total solids. -I have examined a number of waters 
from this section and find the ground water uniformly heavily laden with 
the sulfates of lime and soda with comparatively large quantities of mag¬ 
nesia, especially if the water comes fr'om near the surface as it necessarily 
must in the case of drain waters. In addition to the above constituents, 
there is uniformly some carbonate present which for the present purpose we 
may consider as sodic carbonate. 
We have then an ample supply of these substances present which are 
known to act deleteriously upon cement, i.e. the sulphates, a portion of 
which is magnesic sulfate, and in addition there is so much sodic carbonate 
present that its action cannot be neglected. 
For the present I will assume that the cement used was at least an 
average Portland cement with not more than two per cent, of calcic sul¬ 
fate added. Assuming this to be correct, the action of the waters on the 
tile shows an interesting line of action. 
A preliminary examination of the white decomposition product indica- 
*Taylor & Thompson Concrete Plain and Reinforced P. 401. 
