It was found that in the case of Portland cement, neat, the highest result was obtained with the largest (40) per cent, of screenings, but with one and two parts sand, the strength steadily fell as larger amounts of screenings were used. With Louisville natural cement the presence of screenings seemed to have little effect on neat tests; and with one part of sand to one of cement, the use of as much as 30 per cent, of screenings to 70 per cent, of sifted cement did not appear to decrease the strength. With two parts sand to one cement, the results were slowly diminished by successive additions of larger percentages of screenings.

85. M. R. Feret is said to have replaced with sand the grains of cement retained on sieves having 5,800 and 32,300 meshes per square inch, and found that, except in the case of neat cement mortars, the substitution of sand for coarse particles of cement did not decrease the strength. In experimenting on this subject Mr. Eliot C. Clarke2 found that the coarse particles of cement were somewhat better, for use in mortar, than fine sand, but very little better than coarse sand.

1 Jour. Assn. Engr. Soc, 1890, and Engineering News, Jan. 31, 1891.

2 Trans. A. S. C. E., Vol. xiv, pp. 158-162.

Table 13. Effect Of Removing Coarse Particles From Natural Cement

Fineness Of Cement

Note: — All cement from same barrel, Brand Bn, Sample 27s. Sand, crushed quartz 20-30.

All briquets made by one molder and stored in one tank.

All results, mean of 5 briquets, except two which are means of ten and two briquets, respectively. A — Cement passing No. 20 sieve, holes .033 inch square. B— " " 50 " " .012

C— " " " 100 " " .0065 "

D — " retained on No. 50, reground to pass No. 100. E — " passing No. 50, retained on No. 100.

86. The tests given in Table 13 were made under the author's direction to determine the effect of sifting and the value of coarse particles. It is seen that in neat tests the strength is slightly diminished by sifting out the coarse particles; in the tests of mortars containing equal parts by weight of sand and cement, there is little difference in the strength of the three samples, though the coarser cement appears to gain its strength a little more rapidly. With two parts sand to one of cement, the greater value of the fine particles is very noticeable, and with one-to-three mortars the difference is still more marked, the sifted cement giving 80 per cent, greater strength than the unsifted.

87. In Table 14 these results are arranged in a different way. If we assume that the particles that will not pass the No. 120 sieve are not cement at all, but equivalent to sand, and that all particles passing this sieve are cement, we obtain a new set of proportions of sand to cement. Thus the sample of cement passing No. 20 sieve, sample A, would be composed of 64 parts cement and 36 parts sand, and the 1 to 3 mortar would have in reality the proportion 64 cement to 336 sand, or 1 to 5.2. It is seen that the tensile strength bears a closer relation to the richness of the mortar when considered in this way. There is, of course, no abrupt division in size such that coarser particles act only as sand, while finer ones enter into combination as cement; part of the coarse particles will have some cementitious value, while some of the finer particles will have somewhat the effect of sand.

Table 14. Effect On Tensile Strength Of Removing Ccarse Particles From Natural Cement

Table 15. Value Of Coarse Particles Of Cement, Natural And Portland

Ref. No. 1. Natural cement passing No. 20 sieve. " " 2. Natural cement passing No. 80 sieve. " " 3. Natural cement reground before sifting, until all passed No. 80 sieve.

" " 4. Natural cement, 64 3/4 per cent, of cement passing No. 80 sieve mixed with 35 1/4 per cent, of limestone screenings retained between Nos. 20 and 80 sieves.

" " 5. Natural cement, 64 3/4 per cent, of cement passing No. 80 sieve mixed with 35 1/4 per cent, of crushed quartz retained between Nos. 20 and 80 sieves.

" " 6. Portland cement passing No. 40 sieve.

" " 7. Portland cement passing No. 80 sieve.

" " 8. Portland cement reground (before sifting) until all passed No. 80 sieve.

" " 9. 81 2/3 per cent, of cement passing No. 80 sieve mixed with 18 1/3 per cent, limestone screenings retained between Nos. 40 and 80 sieves. " " 10. 81 2/3 per cent, of cement passing No. 80 sieve mixed with 18 1/3 per cent, crushed quartz retained between Nos. 40 and 80 sieves.

Of the natural cement passing No. 20 sieve, 35 1/4 per cent, was retained on sieve No. 80, while 64 2/3 per cent, passed the No. 80 sieve. In lines 4 and 5 the coarse particles of cement (20-80) were removed and replaced by an equal weight of sand grains, retained between sieves 20 and 80.

Of the Portland cement passing No. 40 sieve, 18 1/3 per cent, was retained on sieve No. 80, while 81 2/3 per cent, passed sieve No. 80. In lines 9 and 10 the coarse particles of cement (40-80) were removed and replaced by an equal weight of sand grains retained between sieves 40 and 80.

All briquets made by same molder, each result mean of five specimens.