is ,2 King: Philippine Concrete and its Aggregates 129 
to determine the mineralogical nature of the grains it was gen- 
erally necessary to free them from adhering dust, dirt, or clay 
by washing with water. Representative pieces of the coarse 
aggregates (gravels and stone) were fractured by rapping them 
sharply with a geological hammer, and the clean broken surface 
was examined macroscopically. 
TECHNIC OF CONCRETE TESTS 
Method of mixing and molding . — All proportions used in this 
paper to indicate concrete mixtures are by volume. For propor- 
tioning the mixtures the weight of a cubic foot of cement was 
assumed to be 90 pounds. Both the sand and the gravel were 
measured dry, and compacted in the measuring boxes by gentle 
percussion. After having found the weight of a unit volume 
of the compacted aggregate it was more convenient to pro- 
portion the mix by using the weight of a unit volume of the 
materials. The aggregates required for a mixture were weighed 
instead of actually measured by volume, and the loss in time and 
the tediousness of compacting them were thus avoided. Like- 
wise, it was more convenient and more accurate to weigh the 
cement for the various tests and then to calculate the volume, 
than actually to measure by volume the cement required. The 
sand and cement were then intimately mixed with a shovel on 
a concrete slab, after which the gravel was added and the whole 
thoroughly mixed in the dry state. In these laboratory tests 
the batches never weighed more than 30 kilograms, so that it 
was a comparatively easy matter to secure a homogeneous mix- 
ture. A crater was next made in the middle of the heap and 
water amounting to from 7 to 10 per cent (depending upon 
the granular sizes of the sand and gravel) of the total weight 
of the dry batch was added; fine aggregates always required 
more water than coarse ones. The consistency sought in these 
tests may be described as “quaking such a mixture is a fairly 
stiff one, which upon slight tamping yields water on the surface. 
The mass does not flow readily and has to be spaded and rammed 
vigorously to produce specimens free from honeycombing. The 
test pieces were made by tamping the concrete into the molds 
in layers about 1 inch thick; the mold was also subjected to a 
vibratory motion by rapping the outside with a wooden block. 
When the mold was full, the top surface of the specimen was 
smoothly finished with a trowel. Most of the laboratory speci- 
mens were 6-inch cubes; some concretes made of small-sized 
gravels were cast in the cylindrical molds already described 
under mortar tests. At the end of twenty-four hours the speci- 
