The preceding paragraph assumes that the measuring is done by hand, but there are two automatic measuring machines on the market which measure the ingredients by volumes loose. One of these machines, the Trump mixer, consists of several bottomless storage cylinders from the bottom of which each of the ingredients flows into a revolving platform from which it is scraped off by stationary arms resting upon the top of the platform and projecting into the material a sufficient distance to scrape off the proper amount of each. The other of these machines, the Gilbreth measurer, consists of several drums, one for each material, placed directly under the storage bins and rotating upon the same horizontal shaft, the quantity of the materials being regulated by means of gates in the bins.
While a statement of the proportions used in practice may be of interest, it can not be of any great value, since it is impracticable, if not impossible, to describe fully the circumstances and limitations under which the work was done. Further, the specifications and records from which such data must be drawn are frequently very indefinite. It is believed that the following examples are as accurate as it is possible or practicable to make them, and also that they are representative of the best American practice.
For foundations for pavements: 1 volume of natural cement, 2 volumes of sand, and 4 or 5, and occasionally 6, volumes of broken stone; or 1 volume of portland cement, 3 volumes of sand, and 6 or 7 volumes of broken stone. Occasionally gravel is specified, and more rarely gravel and broken stone mixed.
For foundations and minor railroad structures: 1 volume of natural cement, 2 volumes of sand, and 3 to 5 parts of broken stone; or 1 part portland cement, 3 parts sand, and 4 or 5 parts broken stone.
For important bridge and tunnel work: 1 part of portland cement, 3 parts of sand, and 4 or 5 parts of broken stone.
For foundations: 1 part portland cement, 1 part sand, and 2 parts broken stone.
For reinforced concrete structures: 1 volume of portland cement, 2 or 21- volumes of sand, 4 or 5 volumes of broken stone.
In harbor improvements the proportions of concrete range from the richest (used to resist the violent action of waves and ice) to the very leanest (used for filling in cribwork). At Buffalo, N. Y., an extensive breakwater built in 1890 by the U. S. A. engineers, con sisted of concrete blocks on the faces and a backing of concrete deposited in place. Portland was used for the blocks and natural for the backing, the proportions being: 1 volume cement, 3 sand, and 84 of broken stone and pebbles mixed in equal parts.
For the retaining walls on the Chicago Sanitary Canal, built in 1895-97: 1 part natural cement, 14 parts sand, and 4 parts un screened limestone.
For the dams, locks, etc., on the Illinois and Mississippi Canal, 1893-98: 1 volume of loose portland cement, 8 volumes of gravel and broken stone; or 1 volume of loose natural cement and 5 volumes of gravel and broken stone.
For the Poe Lock of the St. Mary's Fall Canal, constructed in 1890-95: 1 part natural cement, 14 parts of sand, and 4 parts of sandstone broken to pass a 24-inch ring and not a finch screen. The
broken stone had 46 per cent voids loose and 38 when rammed.
For the concrete blocks used in constructing the Mississippi Jetties, built in 1875-80, the proportions were: 1 part portland cement, 1 part sand, 1 part gravel, and 5 parts broken stone.
For incidental information concerning proportions used in prac tice, see Cost of Concrete, 1 421-23, § 1085, and § 1110.
For an interesting account of a method of determining the proportions of a concrete after it has set in place, see Engineering News, Vol. lix, p. 46, January 9, 1908.
Table 28, page 158, gives the quantities of cement, sand, and stone required for a cubic yard of con crete of different proportions, using three grades of broken stone or gravel. The concrete WAS mixed wet and also mixed very thoroughly. If it had been mixed drier or less thoroughly, it would have been less dense, and consequently less quantities of materials would have been required to make a yard.
Data like that in Table 28 are affected by the fineness of the cement, the fineness and the dampness of the sand, the kind and the coarseness of the stone, the proportions of the several sizes of sand grains and stone fragments, the thoroughness of the mixing, the amount of tamping, etc.; and different experimenters have obtained widely different results. Most experimenters obtain a less quantity of ingredients per cubic yard than in Table 28, probably chiefly because the concrete is mixed drier and entrains more air, and hence is less. dense. For data somewhat similar to that in Table 28, see: Transactions American Society of Civil Engineers, Vol. xlii (1899), p. 109-11, and 137; Johnson's Materials of Construction, p. 610a; Tests of Metals, Watertown Arsenal, 1899, p. 786-87; Report of Chief of Engineers, U. S. A., 1895, p. 2922-31.
• c= number of parts of cement.
s = number of parts of sand.
C = number of barrels of packed portland cement required for 1 cu. yd. of concrete.
S—number of cubic yards of loose sand required for 1 cu. yd. of concrete.
G=number of cubic yards of loose gravel or broken stone required for 1 cu. yd. of concrete.
"If the coarse material is broken stone screened to uniform size, it will contain less solid matter in a given volume than average stone, and hence about 5 per cent should be added to quantities of all three ingredients as computed by the above rule. On the other hand, if the coarse material is well graded in size, about 5 per cent may be deducted from all of the quantities." The above formulas are sometimes modified by changing the constants 11 and 3.8. For example, one engineer substitutes 9.5 for the 11, and 4 for the 3.8.