TIER EROPERILE SIO BUILDING STONES, ETC 353 
Effect of temperature changes.— The durability of a stone 
depends very largely upon its capacity to withstand temperature 
changes. Such changes may affect the mineral constituents of 
rock through expansion and contraction, or they may cause the 
interstitial water to freeze and thaw, on account of which the 
strength of the stone may be materially lessened. 
Very few tests have thus far been made to determine the 
effect on building stone, of the alternate freezing and thawing of 
interstitial water. The importance of such experiments has never 
been questioned, but the difficulty of manipulation and the many 
conditions which need consideration before conclusions can be 
drawn from the quantitative results, have had the effect of almost 
excluding these tests from the experiments on building stone. 
The effect of alternate freezing and thawing may manifest 
itself in three ways: first, cracks may form ; second, small par- 
ticles or grains may be thrown off from the surface occasioning 
a loss in weight; third, the strength of the sample may be 
lessened. The first result is very seldom observed in testing 
samples in the laboratory, owing to the careful selection of the 
pieces tested. The other two, however, usually occur and can 
be measured quantitatively. 
Two methods, known as the natural and the artificial, have 
been employed to determine the effect of alternate freezing and 
thawing of the interstitial water. The natural method is to soak 
the samples with water and alternately freeze and thaw them, a 
- few or many times, at the convenience of the operator. The 
artificial method is to saturate the stone in a boiling solution of 
soluble salt, such as sodium sulphate, and then allow it to dry. 
As the water evaporates the salt crystallizes and expands, produc- 
ing stresses similar to those which result from freezing water. 
It appears to me that the only instance in which it is excusable 
to use this method is when there is no opportunity to freeze the 
samples under conditions which more nearly accord with those 
which occur in nature. 
For the purpose of testing stone according to the natural 
method, the operator should use two-inch cubes as in the 
