EXPANSION" AND CONTRACTION OF CONCRETE. 15 
to insure that the temperature was not changing. The instrument 
then was mounted on the gauge bar, and a jet of steam was passed 
through the inner casing surrounding the steel and brass tubes. 
When the temperature of this inner casing had become constant, 
micrometer readings were taken. A stream of cold air then was run 
through, and again temperature readings were taken at each end of 
the instrument, until it was determined that the inside temperature 
had become constant. Micrometer readings were taken again. Sev- 
eral sets of readings made in this way gave values for the coefficient 
of expansion of the steel tube averaging 0.0000110 per degree centi- 
grade, and for the brass tube 0.00001T9. Knowing the difference in 
readings in the steel and brass tubes at the two measured temperatures, 
and having determined the coefficients of expansion of the metals, a 
curve giving the difference in length of the bars at various tempera- 
tures was plotted readily. When readings were made out on the road 
the temperature of the bars of the instrument was obtainable easily 
by means of the difference in their measured length. 
Bronze plugs spaced 10 feet apart were set in the concrete road. 
A depressed cone formed the top surface, and its center was drill ed 
with a one-sixteenth-inch drill. The top of the plug was protected 
with a brass tube which was set flush with the surface of the road, 
and, except when readings were taken, was filled with putty. Great 
care had to be exercised in setting the plugs as nearly 10 feet apart 
as possible so as, not to exceed the range of the instrument. The 
greatest possible care likewise was taken to keep the holes in the 
plugs clean while measurements were being made. 
In manipulating the instrument one operator was required at each 
end. After setting the supporting points of the blocks into the holes 
drilled in the plugs the brass and steel gauge bars were slid gently 
through the casing a very short distance, slightly jolting the blocks. 
This seemed to settle them into place so that check readings could 
be taken. After each reading the instrument was removed entirely 
from the holes and reseated. The readings were repeated again, and 
an average of three or four more was recorded as the true reading 
for the set. The steel-bar reading then was corrected to what it would 
have been if the bar had remained at constant temperature, and in 
this way changes in temperature of the instrument were eliminated. 
RESULTS OF TEST MEASUREMENTS AT CHEVY CHASE, MD. 
Beginning in September, 1912, a length of experimental road was 
constructed on Connecticut Avenue, beginning at Bradley Lane and 
extending north to Chevy Chase Lake, This road was composed of 
six sections of different varieties, as follows, beginning at Bradley 
Lane: 
Section I. Bituminous concrete, Topeka specification. 
Section II. Bituminous concrete, District of Columbia specification. 
