PORTLAND CEMENT CONCRETE PAVEMENTS. 13 
being less slippery when the pavement is first constructed and is pre- 
ferred by some engineers on that account. Smooth surfaces are more 
generally preferred, except on very steep grades, where it is some- 
times desirable to provide grooves or other comparatively deep mark- 
ings at right angles to the direction of traffic in order to afford a bet- 
ter foothold for horses. Such grooves, however, will cause rapid 
deterioration of the pavement under heavy traffic. 
A satisfactory method of finishing the surface is to use a wooden 
float for smoothing out all template markings (PI. Ill, fig. 1) and 
evening up other slight irregularities. This method of finishing 
produces a surface sufficiently rough for all ordinary grades and pos- 
sesses the advantage of being extremely simple. In using the float 
special care must be exercised to keep the pressure of the hand uni- 
form, in order not to produce irregularities in the surface. Wherever 
a depression occurs it should be filled by adding concrete, and not by 
raking mortar into it with the float. The workmen who do the float- 
ing should be provided with one or more light bridges, which span 
the pavement and which can be easily moved as the work progresses. 
Various sizes of floats are used, and provided they are handled by 
skilled workmen the size is not important. The long float shown in 
Plate X, figure 5, requires less skill on the part of the operators than 
short floats. A suitable design for a finishing bridge is shown in 
Plate XI, figure 1. 
JOINTS. 
It is customary to provide transverse joints at regular intervals 
in concrete pavements, to prevent irregular cracks from being pro- 
duced ; and if the width of the pavement exceeds 20 feet, longitudinal 
joints are also usually provided. Concrete contracts and expands 
with changes in temperature and also with changes in its moisture 
content. It also shrinks or contracts upon setting; and since the 
strength of the concrete is then comparatively low, the tensile stresses 
developed are much more likely to produce cracks than equivalent 
stresses developed in older concrete. It is evident that the greatest 
longitudinal stress which can be developed at any section of the 
pavement, due to contraction, is equal to the weight of the pavement, 
included between the section under consideration and the nearest 
free end, multiplied by the coefficient of friction between the pave- 
ment and the subgrade. Therefore, if contraction joints are spaced 
sufficiently close together to prevent this stress from exceeding the 
tensile strength of the concrete, no cracks should occur. 
If no transverse joints are constructed in the pavement, the length 
of the sections between cracks, judging from such limited data as are 
at present available, will vary from 20 to 150 feet, and depends upon 
the kind of aggregate used, the relative richness of the concrete, the 
