STATE GEOLOGIST. | 27 
When great ultimate strength is required, or when variable strains 
‘occur in a structure, Portland cement should be used. 
MIXING THE MORTAR. 
In preparing mortar the sand and cement should be thoroughly 
mixed dry and then water added and the mass carefully mixed again, until 
the mortar has a proper and uniform consistency to work easily and 
smoothly under the trowel. The strongest mortar for any given sand 
and cement is produced when sufficient cement is used to just fill all the 
voids in the sand with a thin coating of cement over each grain of sand. 
Sands vary greatly in their coefficient of uniformity. Some sands are all 
fine, some all coarse, while many are graded from very fine to very coarse. 
It follows, therefore, that a graded sand requires less cement to make 
a mortar having a given strength, than a sand which has uniform size. 
The voids in the latter sand amount to nearly 50 per cent. of the mass. 
EFFECT OF VARIOUS SANDS UPON THE STRENGTH OF MORTAR. 
Standard quartz sand has about 48 per cent. of voids in it when 
measured dry. Lake sand from Sandusky, Ohio, has about 35 to 37 per 
cent. of voids. Bank sand from Mock’s sand bank northeast of Colum- 
bus, containing considerable clay or loam, has about 33 to 34 per cent. 
of voids. 
A class in civil engineering at the Ohio State University carried on 
extensive tests during the winter term of 1902-03, with several brands of 
cement and the three kinds of sand named above. The characteristics 
of the sands were as follows: 
TABLE 3.—Characteristics of Sands. 
Fineness—per cent. of 
sand remaining on a Baia a Voids Weight 
Kind of Sand. No. 20 | No. 30 | No. 50 ee Peni@ant oe Remarks. 
Sieve. | Sieve. | Sieve. | | 
Crushed Quartz 0 | 100.0 0 0 Bue. AG 83 Clean. 
| | 85 to | 
Lake Sand. dal 16.7 | 26.5 49.1 37 | 108 Clean. 
ea to EF Contains 
Bank Sand. 11.6 37.0 | 33.4 18:0; | "3, | 102 labout'70¢ 
ate Che ONS eee see oh of loam. 
Among the cements tested were the Atlas, Giant and Dyckerhoff 
brands of Portland cement. Seven and twenty-eight day tests were made. 
The results of the 28 day tests are given in table number 4. While the 
table shows quite a variation in the results obtained by the different 
testers, it certainly shows a remarkable uniformity in results when the 
