136 REIBLING AND REYES. 
known usually to occur in the 1: 3 sand briquettes as well. Many 
attribute this drop in tensile strength to brittleness, to unequal 
hardening, to a twisting action of the clips, or to an uneven ap- 
plication of the load, and maintain that it is not due to a dis- 
rupting action in the cement itself. Their final argument in 
this respect is that cement is seldom, if ever, used in tension 
and that the compressive strength does not experience this drop 
if sound cement is used. In fact, many investigations showed 
that in time the set-kiln cements often gave the best results. 
Such observations induced greater conservatism in the demands 
for strength in cement specifications; but, owing to the failure 
of Le. Chatelier and others to ascertain the significance of the 
early gain in strength and the lack of knowledge concerning 
the chemistry of the hardening process itself, there have resulted 
divers conclusions, which more refined methods of manufacture 
and testing served to augment rather than to eliminate. 
A consideration of the mechanical stresses involved in the re- 
sistance of concrete to compression indicates that the compres- 
sive strength of concrete should depend more or less upon 
the tensile strength of the cement used. The strength of 
mortar and concrete in compression is really their strength in 
shearing, and for a granular material the strength in shearing 
would be expected to vary with the strength in tension. The 
tensile strength of the cement is also related to a great extent to 
the factors which bind the aggregates together, and which tend 
to prevent disruption and wedging of the coarser copards when 
concrete is subjected to compression. In spite of this, it appears 
that no definite relationship has been established between the 
two. However, it finally has been shown that the compressive 
strength may not develop a uniform, nor a uniformly reg- 
ulated, increase between different periods of time. In fact, the 
compressive strength of sound Portland cement mortar and con- 
crete often shows a drop as decided as that of the tensile 
strength. 
The early failure universally to recognize the truth in this 
respect can be partly accounted for by the difference in the size 
of the specimens used for tension as compared with those used 
for compression. For an exact relationship, it is essential that 
the size of the specimens should always be the same. The outer 
and inner parts of mortar and concrete specimens do not harden 
at the same rate. Chemical analysis has repeatedly shown this 
@ Bull. U. S. Geol. Surv. (1908), No. 331. 
