Table 17 gives the results of a series of experiments to determine the effect of the size of grains of sand upon the tensile strength of cement mortar. The briquettes were all made at the same time by the same person from the same cement and sand, the only difference being in the fineness of the sand. The table clearly shows that coarse sand is better than fine. Notice that the results in line 4 of the table are larger than those in line 3. This is probably due to the fact that the sand for line 4 has a greater range of sizes and conse quently fewer voids. If this explanation is true, the coarse sand is relatively better than appears from Table 17, since the sand in each line of the lower half of the table has greater variety of sizes than those in the upper half. • Table 18 shows the fineness of natural sands employed in actual construction; and as the sands were to all appearances of the same character, this table also shows at least approximately the effect of fineness ;Ton tensile strength. This table agrees with the pre ceding in showing that the coarser sand makes the stronger mortar. This conclusion is perfectly general.
If the voids are filled with cement, uniform coarse grains give greater strength than coarse and fine mixed; or, in other words, for rich mortar coarse grains are more important than small voids. But if the voids are not filled, then coarse and fine sand mixed give greater strength than uniform coarse grains; or, in other words, for iean mortar a small proportion of voids is more important than coarse grains.* Specifications seldom contain any numerical requirement for the fineness of the sand. The two following are all that can be found. For the retaining-wall masonry on the Chicago Sanitary Canal the requirements were that not more than 50 per cent should pass a No. 50 sieve, and not more than 12 per cent should pass a No. 80 sieve. For the Portage Dam on the Genesee River, built by the New York State Engineer, the specifications were that at least 75 per cent should pass a No. 20 sieve and be caught on a No. 40.
The fineness of the sand employed in several noted works is as follows, the larger figures being the number of the sieve, and the smaller figures preceding the number of the sieve being the per cent retained by that sieve, and the small number after the last sieve number being the per cent passing that sieve: Poe Lock, St. Mary's Fall Canal, b 20 15 30 ss 40'b; concrete for pavement foundations in the City of Washington, D. C., ° 3' 6 8 8 1s 10'° 20 82 40 Gen esee (N. Y.) Storage Dam, '20 8 30 " 100'; Rough River (Ky.) Improvement, 11 30 bb 50 St. Regis sand, Soulanges Canal, Canada, 1z 30 Grand Coteau sand,t Soulanges Canal, Canada, 1' 20 so 30 27 50 Tables 18 (page 89) and 19 (page 93) show the fineness of a number of natural sands employed in actual work.
The specifications proposed by the German Concrete Society for concrete structures designate as sand any grain of natural sand or crushed stone less than 0.28 inch in diameter, material having
pieces larger than this being called pebbles or broken stone. Many American engineers draw the line between sand and pebbles or broken stone at pieces 0.20 inch in diameter. A noted American authority says a good sand should have all of its grains less than 0.20 inch in diameter and should have not less than 30 per cent that will pass a No. 40 sieve.* Voids. The smaller the proportion of voids, i.e., the inter stices between the grains of the sand, the less the amount of cement required, and consequently the more economical the sand.
The per cent of voids in sand may be determined by either of two methods, which for brevity will here be designated as (1) deter mining voids by the specific gravity method, and (2) determining voids by direct measurement.
Determining Voids by Specific Gravity Method. This method consists in determining (1) the specific gravity of the sand and from that computing the weight of a cubic unit of the solid material, and (2) the weight of a cubic unit of the sand. The difference between the first weight and the second weight, divided by the first weight, gives the proportion of voids, or expressed in percentages, gives the per cent of voids.
The specific gravity of siliceous sands is quite uniformly 2.65; but glacial sand containing fragments of limestone, sandstone, shale, and slate may have a specific gravity of 2.60 or even a little less. However, it is sufficient to assume the specific gravity of good siliceous sand at 2.65. The sand should be dried at a temperature not less than 212° Fahr. until there is no further loss of weight. The weight of a unit of sand may be determined for the sand loose, shaken, or rammed, and the per cent of voids will be for the corre sponding condition. It is probably better to determine the voids for the sand when rammed, since the mortar is either compressed or rammed when• used.
Determining Voids by Direct Measurement. The proportion of voids may be determined by filling a vessel with sand and then determining the amount of water that can be poured into the vessel with the sand. The quantity of water poured into the sand divided by the amount of water alone which the vessel will contain is the proportion of voids in the sand. The quantities of water as above may be determined either by volume or by weight. The proportion of voids may be determined for the sand loose, shaken, or rammed, the latter condition being the more appropriate, since the mortar is either compressed or rammed when used. For accurate work the sand should be dried to expel all moisture, as moisture affects both the weight and the volume of the sand, particularly if the sand is fine (see 4 196). Further, even though the sand is so coarse that its volume is not appreciably affected by the moisture present, the sand should be dried since it is the total per cent of voids in the sand, and not of the air-filled voids alone, that is desired.