LOCKS, 165 



part of sand. One part of the resulting product was then mixed with approximately 

 three parts of sand and six parts of stone, the proportions thus being i to 19 on the 

 basis of the cement alone. The grinding plant was owned and operated by the United 

 States. 



All the above proportions were by measure, except that in some cases the cement 

 was taken as packed, in others as loose measure. Specifications should always state 

 whether the cement is to be measured loose or packed. The contents of a barrel 

 when shipped vary with different brands from 3.03 cubic feet to 3.35 cubic feet, and 

 when loose from 3.75 cubic feet to 4.19 cubic feet, according to the fineness of the 

 grinding. Tests made from a number of different brands gave an average of 3.18 cubic 

 feet in the barrel, and 4.07 cubic feet loose, the ratio thus being i to 1.28. 



Sometimes a natural cement is used for the parts below water, for the sake of 

 economy, but Portland cement should always be used above. 



The sand should of course be clean, and, where obtainable, of coarse grains, and 

 a good average quality will be obtained by having it fill the following specifications 

 for fineness: To pass a No. 30 sieve, not over 70 per cent; to pass a No. 50 sieve, not 

 over 50 per cent; to pass a No. 100 sieve, not over 2 per cent.* 



The stone may be either broken stone or gravel, and should be free from dirt. 

 For the former limestone is best, as it appears in course of time to combine chemically 

 with the cement, but any good hard stone will answer. If the stone is soft, it is liable 

 to crush under the rammers, besides being deficient in tensile strength. Certain classes of 

 sandstone, however (especially those impregnated with iron), which are soft when quarried, 

 will become reasonably hard after some weeks' exposure to the weather, and can then 

 be used, where better stone is not obtainable, with fairly good results. 



Our experience has led us to conclude that a denser and more uniform concrete 

 is obtained where the stone or gravel does not exceed i inch or i i inches in diameter, and 

 includes all smaller pieces down to J-Q of an inch in diameter. Where stone of 2 or 2 

 inches in diameter is used it bunches together and leaves voids, which with the smaller 

 stone are much less frequent, since it packs closer under the rammers. 



It is occasionally impracticable to secure sand or stone that can be made entirely 

 clean, and in such cases a small amount of earth or foreign matter need not be a drawback, 

 as it will have no appreciable effect in the mass of concrete. Some specifications limit 

 this amount to 2 per cent, f Similarly the gravel will usually retain more or less 

 sand after screening, the variation in the amount being usually limited to 6 per cent. 



* In certain localities it has not been possible to secure a coarse sand without great expense, and 

 in such cases common river sand has been used, and has given results amply strong for all practical 

 purposes. 



f Experiments by the authors with some briquettes composed of one part of Portland cement to one 

 part of sand and river mud, and with others composed of one part of Portland cement to three parts 

 of sand and river mud, showed that where the mud did not exceed five per cent of the amount of sand 

 >io appreciable diminution of strength occurred, and it was not until the amount exceeded ten per cent 

 'iat any marked reduction was observed. 



