LIMITING STRENGTH OF ROCKS UNDER STRESS 131 
diminish with load, while « tends to the value 4. Under these 
ai eee . ae G—l. 
conditions it is quite possible that the condition B< —— given 
eo 
in (22) may be satisfied before the specimen actually begins to rup- 
ture in the neighborhood of the cylindrical cavity. The break- 
down of a rock specimen under these conditions would be charac- 
terized by a tendency to split across a plane through the axis: over 
a cross-section a system of spiral cracks cutting the cylindrical 
cavity at an angle would make their appearance and long splinters 
parallel to the axis would break off from the interior of the cylin- 
drical cavity. Actual cases of fractures of this kind are described 
by Dr. Adams in the preceding paper.t. We note that for a solid 
cylinder (a=o) the stress-difference (i) = (iii), so that the conditions 
for the appearance of the surfaces of shear (i) and (iii) are identical. 
System (iii) is characterized by the appearance of Luder’s lines over 
the surface of the specimen; these in fact are often observed. 
$6. ON THE LIMITING STRESS-DIFFERENCE IN THE NEIGHBORHOOD 
OF A SMALL CYLINDRICAL CAVITY 
The preceding analysis may be applied to a determination of the 
limiting stress-difference S in the neighborhood of the small cylin- 
drical cavities in the specimens of Westerly (Rhode Island) granite 
tested by Dr. Adams and described in the preceding paper. This 
granite gave a crushing strength of 27,370 pounds per sq. in. for 
a two-inch cube. If we note that for this granite >= .2109,7 the 
ae PT 4 : 
condition B>% ~ is satisfied and the greatest stress-difference 
lees . 
in the specimen is P, so that S=P. Experiments 373 and 357 
described in Dr. Adams’ paper (p. 113) enable a limit to be assigned 
to the value of S. In experiment 373 a pressure P of 160,000 
pounds per sq. in. applied during 75 days produced no change in 
the diameter of the cylindrical cavity, indicating that no flow or 
permanent set had taken place. In the case of a similar specimen 
tested under a pressure P of 200,000 pounds for the same length of 
time, it was found that the vertical cavity was partly closed up by 
t Adams, p. 109 of this journal, Plate I, figs. c and d; Plate I], fig. c. 
2 Adams and Coker, loc. cit. 
