EXPANSION AND CONTRACTION OF CONCRETE. 3 
According to these figures, neat cement shrinks about 2J inches in 
100 feet when kept in air six months, and in four years the shrinkage, 
due merely to the drying out of the moisture, amounts to almost -t 
inches in 100 feet. Mortar mixed in the proportions of 1 part of 
cement to 3 parts of sand, when allowed to harden in the air, shrinks, 
according to White's results, from 0.08 per cent to 0.1 per cent. 
These figures confirm in a general way the previously published 
figures of Bauschinger and others. 
As neat cement and mortar were known to have this physical 
characteristic of expanding and contracting, depending upon their 
moisture content, it was considered reasonable to suppose that con- 
crete, too, would show the same general behavior. The following 
laboratory tests were made on concrete to determine the amount of 
movements produced by the drying out and absorption of water. 
These movements, combined with temperature movements, probably 
account for most of the expansion and contraction of concrete pave- 
ments, and a knowledge of their values is of assistance in the design 
of expansion joints. 
DESCRIPTION OF EXPANSION AND CONTRACTION SPECIMENS. 
White and the earlier investigators made tests on the change in 
length of neat cement and mortar with specially designed micro- 
meters on very small specimens, White's being only 4 inches in length. 
Obviously it would be impossible to use such small test pieces as these 
to determine the expansion and contraction of concrete containing 
coarse aggregates, and the type adopted was a column 8 inches square 
and 5 feet high. The column form of specimen was used to permit 
the free expansion of one end, in this way preventing friction from 
affecting the accuracy of the measurements. 
PREPARATION OF SPECIMENS. 
All concrete mixtures were proportioned by weight with a stand- 
ard brand of Portland cement, Potomac River sand, and crushed 
gneiss or gravel. The cement was normal and passed the specifica- 
tions of the American Society for Testing Materials. The sand was 
coarse and clean and considered a good grade of concrete sand. All 
concrete was hand mixed and was made of various consistencies. The 
specimens of very dry consistency required hard tamping to. consoli- 
date the concrete in the mold, while those of wet consistency were 
made simply by puddling the mixture in place. The amount of 
water used for the dry mixtures was about 8.5 per cent of the weight 
of dry materials. For the wet mixtures from 10 to 12 per cent was 
used. 
