164 REIBLING AND SAIJNGER. 



flat square on the slab, and by pressure with the edge of the trowel 

 rapidly divide it into sixteen cubes. One such cube forms each layer 

 of the briquettes. 



In tamping the last layer, it is advisable first to lay an empty mold 

 exactly over the other. The empty mold acts as a guide for the tamper 

 and so avoids the possibility of the loss of the full effect of a blow caused 

 Ijy striking the edge of the mold. 



The Fnited States Army specifications direct that each layer of cement 

 in the molds be uniformly tamped with thirty blows. There is no 

 possible way to avoid unequal overlapping of blows with the tamper 

 specified (both round and square). As a result, and also because of 

 the fact that the cement is put into the molds in a more or less lumpy 

 condition, certain voids and excess in the consistency and compactness 

 of the resulting briquettes are unavoidable. Air spaces also form with 

 more or less irregularity. These produce internal strains and variation 

 in cohesion, and consequently differences in the breaks. This illustrates 

 one case of a sjiecification which imposes variability of results upon the 

 tester. 



The American Society method eliminates the greater part of this 

 trouble. The paste is more homogeneous and plastic (not lumpy) ; it 

 is readily pressed into the molds by the fingers and a subsequent patting 

 of the briquette with the flat side of the trowel will eliminate any varia- 

 tion in compactness caused by unequal pressure of the fingers. 



The natural tendency in tamping briquettes is to strike the middle, 

 narrow section more than the wider ones; it follows that the resulting 

 briquette is denser in the middle portion. This is the main cause of 

 bad breaks, besides giving a higher result than is just if uniformity 

 of tamping is followed. It is just as essential not to weaken the middle 

 section below the average density. Such a method of tamping will give 

 good breaks, but lower the tensile strength. 



After experimenting with many methods to secure as uniform tamping 

 as possible, conducive to good breaks and greatest strength, we have 

 adopted the following method which can be accurately carried out with 

 our automatic tamj)er. 



Fig. 11. 



