xv, i Witt: Portland Cement Raiv Mixture 115 
not be eliminated. There is another class of errors due to the 
composition of the substances under investigation; errors of 
this class have been found to be largely responsible for the non- 
concordant results of Table II. 
In the first column, the calcium is determined, and the results 
are calculated to calcium carbonate. In the case of the calci- 
meter or with any other method in which the carbon dioxide is 
determined, the result in terms of calcium carbonate must be 
calculated from the total carbon dioxide found. This, of course, 
assumes that all calcium is present as carbonate and that all 
carbon dioxide is combined with calcium, which in this case is 
not true. Carbonates other than calcium are present in small 
amount in the limestone, causing the results to be slightly too 
high. However, the calcium in the two siliceous materials is 
not present as carbonate and consequently is not included when 
the determination by calcimeter is made. The net result is that 
the values obtained are low. 
In the acid-alkali titration method, the alkalinity or acid- 
consuming power is calculated to calcium carbonate. In this 
case the errors due to apparatus and to manipulation may be 
made very small. However, there are much greater errors due 
to the composition of the mixture. On being treated with 
standard acid, all carbonates go into solution and are calculated 
to calcium carbonate. Of the calcium in the siliceous materials, 
a portion goes into solution without neutralizing acid equivalent 
to that which would be required for calcium carbonate, and a 
portion is undissolved by the acid. Some of the iron and the 
aluminium in these materials go into solution and consequently 
neutralize the acid, thereby tending to produce a higher result . 37 
The results obtained by determining the acid-insoluble matter 
were so irregular that no further analyses by that method were 
made. 
Slight errors are caused by the presence of moisture in the 
raw mixture as it comes from the tube mill. It has not been 
found practicable to eliminate these in the routine work. The 
moisture content of the material as it enters the tube mill is 
determined at regular intervals and is kept as low as possible 
37 S. B. Newberry, Cem. and Eng. News, March (1903) 35, describes a 
double titration method by which correction for magnesium may be made. 
With the raw mixture under investigation, however, the error due to the 
effect of iron and aluminium cannot be corrected. The error due to mag- 
nesium is not corrected because it tends to compensate the other one. 
