TESTS ON CEMENT 211 



may be due to incorrect proportioning or to insufficient grind- 

 ing of the raw materials, to underburning, or to lack of 

 sufficient storing before use, called seasoning. A certain 

 amount of seasoning is usually necessary, because almost 

 every cement, no matter how well proportioned or burned 

 it may be, will contain a small amount of this excess of lime, 

 which, on standing, will absorb moisture from the air, slake, 

 and become inert. 



Excess of magnesia or the alkalies may also cause unsound- 

 ness, but the ordinary cement rarely contains a sufficient 

 amount of these ingredients to be harmful. Sulphate of 

 lime is occasionally responsible for unsoundness, but this 

 ingredient usually acts in the opposite direction, tending to 

 make sound a cement that otherwise might disintegrate. 



The property of soundness is determined in one or more 

 of three ways: by measurements of expansion, by normal 

 tests, and by accelerated tests. 



Measurements of expansion are made by forming specimens 

 of cement, usually in the shape of prisms, and measuring the 

 change in volume by means of a micrometer screw. At the 

 present time, however, it is believed that expansion is not a 

 sure index of unsoundness, so that this test is seldom 

 employed. 



Normal tests consist in making specimens of cement mixed 

 with water, preserving them in air or in water under normal 

 conditions, and observing their behavior. The common 

 practice is to make from a paste of neat, or pure, cement on 

 glass plates about 4 in. square, two circular pats, about 3 in. 

 in diameter, i in. thick at the center, and tapering to a 

 thin edge. These pats are kept in moist air for 24 hr.; 

 then one of them is placed in fresh water of ordinary tempera- 

 ture and the other is preserved in air. The condition of the 

 pats is observed 7 da. and 28 da. from the date of making, 

 and thereafter at such times as may be desired. 



The most characteristic forms of failure are illustrated in 

 Figs. 1 and 2. 



Fig. 1 (a) shows a pat in good condition. 



Fig. 1 (6) illustrates shrinkage cracks that are due, not to 

 inferior cement, but to the fact that the pat has been allowed 



