XIII, A, 4 Witt and Reyes: Solubility of Portland Cement 151 



MANIPULATION 



Since it was likely that the larger particles of cement were 

 the last to be affected by the water, these were eliminated be- 

 fore starting the second series. An air separator essentially 

 similar to the Goreham flourometer " was utilized. No attempt 

 was made to obtain quantitative results nor to measure the size 

 of grain. The air pressure corresponded to 20 millimeters of 

 mercury. The air was passed through suitable solutions to re- 

 move both moisture and carbon dioxide, before coming into 

 contact with the cement. Since cement dust is likely to be 

 slightly different in chemical composition from the original 

 cement after such a separation, the cements were again analyzed, 

 with the results shown in Table III. All the work hereafter 

 described was done with this material. 



Table III. — Analyses of cements after air separation. 



[Numbers indicate percentages.] 





Brand. 



I. 



n. 



III. 



IV. 



Loss on ignition 



Silica (SiOi) 



3.91 

 20.48 

 7.81 

 2.37 

 61.14 

 1.22 

 1.07 

 1.97 



3.73 

 20.22 

 7.17 

 2.11 

 62.38 

 . 1.30 

 0.64 

 2.43 



3.55 

 18.96 

 9.58 

 2.32 

 61.84 

 1.42 

 0.63 

 1.62 



5.00 

 18.40 

 8.95 

 2.03 

 62.20 

 1.34 

 0.87 

 1.27 



Alumina (AlaOa) 



Iron oxide (Fe203) 



Calcium oxide (CaO) - 





Sulphuric anhydride (SOa) 





Some other changes also were found advisable before starting 

 the next series of determinations. It was found that the Erlen- 

 meyer flasks did not stand the continued rough usage in the 

 shaking machines. It was also desirable to increase the actual 

 amount of water for each experiment as well as the quantity 

 per gram of cement. Therefore the new manipulation was as 

 follows : 



A 20-liter bottle was filled with water free from carbon dioxide 

 and protected by a soda-lime bulb. A special automatic pipette 

 was made with an approximate capacity of 850 cubic centimeters. 

 When this was standardized, it was found to deliver 863.5 cubic 

 centimeters. This value was constant and was sufficiently close 

 to the desired volume, so it was not changed. The pipette was 

 mounted and then connected with a siphon in the 20-liter bottle. 



"Cf. Tech. Paper, U. S. Bur. Standards (1915), No. 48, 8. 



