6 BULLETIN 249, IT. S. DEPARTMENT OF AGRICULTURE. 
cured from gravel containing a considerable percentage of sand, 
or from crushed limestone, a f-incli mesh minimum screen is to be 
preferred. 
Water used in mixing concrete should be reasonably clear and free 
from alkalies, acids, vegetable matter, or other injurious materials. 
The subject of water supply will be later discussed under the heading, 
" Methods, organization, and equipment." 
PKOPOBTIONING. 
Concrete in pavements is subjected to much more severe service 
conditions than that in walls, foundations, etc. Most of the old 
rules for proportioning concrete were developed with a view to pro- 
viding only for simple compressive stresses, such as are met with in 
the latter class of structures. Hence it is not surprising that the 
early results obtained for pavements by following the old rules were 
not generally satisfactory. Concrete pavements must resist not only 
crushing and impact stresses but the wearing action of traffic as well, 
and this is probably the most destructive process to which they are 
subjected. 
The essential qualities which enable any material to withstand the 
wearing action of traffic are hardness and toughness. Laboratory 
tests have been devised for determining the relative degree in which 
these qualities are possessed by different kinds of stone and brick, but 
none of these tests is suitable for making similar determinations re- 
garding concrete mixed in different proportions and composed of 
different materials. The reason for this is that the structure of con- 
crete, unlike that of ordinary stone and brick, is not homogeneous. 
It is possible, however, to employ the routine road-material tests 
described in Office of Public Roads Bulletin No. 44 on the mortar and 
coarse aggregate separately, and it would seem that the results which 
might be obtained in this way ought to furnish a fairly reliable 
index to the quality of concrete which could be produced from the 
materials tested. The proper proportions in which to mix the ma- 
terials can probably be best determined from actual service tests. 
* Plates VII, VIII, and IX are diagrams showing the relative hard- 
ness, toughness, and crushing strength of mortars mixed in different 
proportions and in which two different qualities of sand were used. 
Sand for one set of the test specimens, as noted on the diagrams, was 
standard Ottawa, while that for the other set was natural quartz 
sand which showed the following analysis: 
