is probable that the silicates also recombine with some of the water. This is the first step, and produces the so-called setting. It is best passed through while the cement is exposed to the air, and is the reason why cement mixtures must be used as soon as moistened. But, this now complete, a more complex process is set up. The moistened cement brought in contact with the air, or exposed to water, at once begins to absorb carbon dioxide, for all ordinary air contains the gas, and most water holds it in solution. The gas unites with the lime to form a carbonate again, and this goes on until the whole of the lime is turned back to limestone. The same change occurs in the magnesia, but in this the action proceeds more slowly. With a pure lime cement this action is probably nearly complete at the end of a few months; but, with a cement containing magnesia, it will continue for many years. The strength of the cement increases so long as the change continues. So a Portland cement will develop its full strength in a few months, while our natural cements will not for years, and, so long as it continues, the structure improves.
Rough testing of cement, so as to enable a workman to get a crude and imperfect idea of its value, is easy. Enough of the pure cement should be taken to make a ball an inch in diameter and mixed with just sufficient water to make it mold readily and be rolled into a ball. Then it should be exposed to the air and left for two hours. At the end of that time it should be set; then it should be put into water and left. It should grow gradually harder, and should show no signs of cracking or crumbling, even when left for ten days. Any cement that does not endure this test is not of sufficiently good quality to make satisfactory structures; any cement that stands this properly will be generally satisfactory if properly used.
In determining how to construct a building, a series of tests is often required that shall show tensile, breaking, twisting, and crushing strength, and also adhesion of the materials used for mortar. No one of these can be dispensed with, since material that will endure one satisfactorily will often fail utterly in another, and hence prove worthless for the use desired; but for general purposes the test of cement which is the most valuable is that which determines its tensile strength. Comparative tests of this show the value of cements from different sources better than any other one test.
To make an accurate test of any lot of cement, great care is necessary in selecting and manipulating the samples. The test sample ought not to be taken from a single package, but from several in equal quantities and thoroughly mixed. The sample must also be carefully protected from air and moisture until the test is made. When used, it must be molded with just the right