EXPANSION AXD CONTRACTION OF CONCRETE. 7 
concrete. This was done by covering them with burlap and keeping the bur- 
lap wet constantly. As soon as the concrete had become sufficiently hard, or 
in most instances not over two days after pouring, the forms were removed 
and initial readings of the specimens were made. 
Referring to the curves, which are plotted with age in days as abscissas and 
with unit expansions and contractions as ordinates, note that during the first 
15 days when the specimens were kept wet there was a continuous expansion,' 
the maximum amounting to about 0.0001 inch, or 0.01 per cent. At the end of 
15 days the specimens were permitted to begin drying out, and the effect of this 
drying is seen on the curve, which drops, or shows contraction, immediately. 
The specimens continued to contract for a period of practically one year, when 
the maximum contraction was about 0.0006 inch, or 0.06 per cent. After the 
specimens were more than a year old (460 days and 540 days) they again were 
kept continually moist and immediately started to expand, but did not regain 
their former length. The amount of expansion was roughly 0.0004 inch per 
inch of length, or about 0.0002 inch short of their original length. Note that 
during the first few days after reimmersion these specimens expanded only 
from 0.0001 to 0.0002 inch per inch of length and their subsequent expansion 
was slow, requiring several months to reach the maximum. The above speci- 
mens were dried in the rather dry and warm atmosphere of the laboratory, and 
therefore were almost as dry as it was possible to make them. Tt will be 
noticed that there is no great difference in the contraction and expansion of 
the different mixtures tested, whether they were 1:2:4 or 1:3:6, very wet or 
very dry. Comparing figure 5 with figure 6, it will be noted that the ultimate 
contraction is about the same whether or not the concrete was subjected to an 
initial period of wetting. 
Let it be supposed that the ends of a concrete construction are immovable 
and that it is subjected to extremely dry conditions, so that it will shrink 
0.0006 inch per inch of length. Then, if a modulus of elasticity of 3,000,000 
pounds per square inch is accepted as correct, the tensile stress produced will 
be 3.000.000 X0.0006=1S00 pounds per square inch. Obviously, under such con- 
ditions as these, concrete must crack, since a stress of only 200 pounds per 
square inch is not far from the maximum tensile strength of concrete. There 
are, however, very few concrete structures that are subjected to the degree of 
drying out suffered by the specimens reported, and the results are given merely 
to show what can and does happen to concrete under conditions favorable to its 
thorough drying. 
These curves show quantitatively what occurs in a concrete pavement just 
after it is laid and demonstrate the efficacy of the practice of keeping the 
concrete wet for a short period after pouring. The curves show that concrete 
remains expanded as long as it is wet and contracts as soon as it begins to 
dry out, with the consequent tendency to form cracks. In a concrete pavement, 
where there is always some restraint due to friction at the base, any tendency 
toward shrinkage will be resisted, and tension will be developed. In the 
earlier stages of hardening the tensile strength of concrete is very low, and 
consequently a very minute shrinkage may produce cracking, irrespective of 
the presence of expansion joints. 
A consideration of the excessive shrinkage of concrete due to complete 
drying out, with a resulting hazard of overstress in tension, shows the great 
necessity for maintaining concrete structures in a moist condition for a few 
weeks after pouring. By this practice it is seen that they may be kept 
expanded, and, therefore, under small compressive stress. When they do dry 
