PHYSICAL AND CHEMICAL PROPERTIES. PARTS IV, V. 137 
to be true, but in many cases the difference is perceptible even 
to the eye. Most cement testers have noticed the inner core of 
almost dry mortar in freshly broken briquettes. The size of 
this core varies with the age, permeability, and nature of the 
cement, and the penetrating power of the reacting substances. 
For any given penetrating effect the value of the ratio of the 
zone influenced to the remaining interior decreases as the 
volume increases. Consequently, large specimens are less apt 
to show the peculiarities of cement than small ones. Also, as a 
retrograde movement in the cement near the surface may be 
counteracted by a development of strength in the inner ma- 
terial, and vice versa, we find that 2.54-centimeter specimens 
are more apt to give compressive tests showing the up and down 
values characteristic of 2.54-centimeter tensile briquettes than 
the 5.08-centimeter cubes so universally used. 
The size of the specimens also influences the value of the ratio 
between the tensile and compressive strengths. For instance, 
when 5.08-centimeter cubes were used, all writers agreed that 
the value of this ratio was between 8 and 12 for 1:3 mortars. 
However, W. Pruves Taylor™ recently conducted a long series 
of tests in which he obtained compressive strengths of small 
cylinders 2.54 centimeters high and having a diameter giving 
an area of 6.45 square centimeters. These tests gave ratios the 
values of which averaged but slightly over 4 for the ordinary 
1:3 mortar with quartz sand. Taylor says that— 
another frequently made statement that tests in compression show a 
uniform increase in strength between different periods is not borne out 
by these tests, the strength in compression showing the common up and 
down values of sufficient frequency and amount to preclude the possibility 
of being caused by error. 
However, as Taylor further remarks, it still appears that— 
the values do not go up and down together, showing that not even for a 
single test can the ratio be considered a constant quantity, and that tensile 
and compression strengths are really entirely distinct properties and are 
only interdependent to a slight degree. 
Our experience in crushing the halves of the briquettes broken 
in tension has also demonstrated the up and down values of this 
test. Sometimes the changes in compression coincided with the 
changes in tension and very often the direct opposite was true; 
but here, too, we have unlike conditions, the area of the cross 
section of the broken halves being considerably larger than the 
* Amer. Soc. Testing Mat. (1906), 6, 389. 
