SHRINKAGE IN ROLLING. Before beginning to spread the layers of stone, it is necessary to inquire as to the amount the crushed stone will shrink in rolling. The shrinkage has an impor tant bearing upon the thickness and cost of the finished road; but the amount of shrinkage is often greatly over-estimated. It is fre quently stated that rolling will cause broken stone to shrink 33 per cent; or, as it is usually put, 6 inches will roll down to 4. Ap parently this statement is based upon MacAdam's experience; but MacAdam used neither a binding material nor a roller, and the road was compacted only after months of travel, when the traffic had pulverized sufficient material to bind the road and after much of that which had been pulverized had blown away. The follow ing examples from actual practice show no such shrinkage.
In one case,* with trap rock 1* to 21 inches, rolled with a 12* ton steam roller upon a subgrade so hard that the wagons hauling the stone made no ruts, 5.67 inches of loose stone rolled to 4 inches, and 7.3S inches rolled to 6 inches. The average thickness of the loose stone was determined by dividing the quantity of stone used by the area covered. The first is a shrinkage of 29 per cent and the second of 19 per cent. The difference between these two re sults is probably due to errors of observation, to variations in the thickness of the finished road, and to the fact that the thicker layers did not compact as solidly as the thinner ones. The stone was rolled dry until the desired thickness was reached, when the binder was added, and sprinkling was commenced.
In another with 2-inch trap laid on the compact surface of an old crushed-stone road and rolled with a 12-ton roller, 3.9 inches of loose stone rolled to 3 inches. The shrinkage was 23 per cent. The thickness was determined from the area covered and the quantity of stone used. No stone could have been forced into the subgrade, but there was some uncertainty as to the average elevation of the surface of the old street.
It has been determined / by tests over several miles of road where the output of the crusher was carefully measured in wagons and also when rolled in place, that 6 inches of loose hard limestone will roll down to 4I inches, which is a shrinkage of 20 per cent.
Itis probable that the maximum actual shrinkage in roll ing is less than 20 per cent. The apparent shrinkage depends upon the nature and condition of the subgrade, i. e., upon the amount of stone forced into the earth.
If the soil is clay, the sprinkling required to work the binder into the interstices may soften the subgrade so that considerable stone will be forced into the earth. This condition is indicated by the roller's leaving tracks upon the surface; and when this occurs, the work should be stopped until the subgrade dries out. To prevent
the crushed stone from being forced into the clay subgrade during construction or after completion—particularly when the frost is going out,—a layer of sand, stone screenings, ashes, or the like, is sometimes interposed. The English engineers often use " hard core" (a mixture of brick rubbish, old plastering, and broken stone) on a clay soil, to prevent the mud's working into the metaling. Any material not affected by water is useful for this purpose; and the finer it is the better, since the smaller will be the apertures in it, and the more certainly will it prevent the soil from coming up through it.
If the soil is sandy, a thin _ayer of coarse gravel or broken stone laid upon the surface and then rolled, will prevent any further loss of the road metal in the subgrade. If the soil is nearly pure sand, the wetter it is the less crushed stone will be forced into it; and therefore if water is plentiful, it may be wise to keep the sand satu rated while the rolling is in progress to prevent the loss of the stone. The Massachusetts Highway Commission used cotton cheese-cloth on a soft fine sand to prevent the stone from sinking into the sub grade. " It is not at all needful that the partition should be endur ing, for as soon as the stones in the lower layer have been forced into contact and have become bound together, there is no further danger of the mingling of the stone with the sand; and hence the speedy decay of the fabric is a matter of no consequence. The cloth was spread in strips lengthwise of the way; and the stone for the bottom layer was shoveled from the sides upon it with no unusual care. A section through such a road shows that the stones do not tear through the cloth. At 3 cents per square yard on the road, the cost of the cloth may be less than one third that due to the loss of the broken stone which would occur if it were allowed to come directly in contact with the sand. Various kinds of strong paper were tried, but found worthless." * "A thick coating of straw has been used to hold up the macadam on a sandy soil."t However, if the sand is firm enough to hold up the stone during the rolling, it is not necessary to prevent the mixing of the sand and the stone, since the subgrade may be left a little high, with the expectation of forcing the stone into the sand. This is equiv alent to using the sand of the subgrade as a filler or binder for the lower portion of the broken stone. If the sand is dry and nearly pure, it can be thus forced nearly to the top of a 4-inch course of coarse broken stone.