THE PHENOMENA OF RUPTURE AND FLOW IN SOLIDS. 
181 
almost perfectly elastic up to these high stresses. The strength diminished, however, 
as time went on, until after the lapse of a few hours it reached a steady value whose 
magnitude depended on the diameter of the fibre. 
Similar phenomena have been observed with other kinds of glass, and also with fused 
silica. 
The relation between diameter and strength in the steady state was investigated in the 
following manner. Fibres of diameters ranging from 0 • 13 X 10 -3 inch to 4 • 2 X 10 -3 inch, 
and 6 inches long, were prepared by heating the glass to about 1400° C. to 1500° C. in 
an oxygen and coal-gas flame and drawing the fibre by hand as quickly as possible. 
The fibres were then put aside for about 40 hours, so that they might reach the 
steady state. The test specimens were prepared by breaking these fibres in tension 
several times until pieces about 0 • 5-inch long remained ; these were then tested by 
the balance method already described. The object of this procedure was the 
elimination of weak places due to minute foreign bodies, local impurities and other 
causes. 
Table V. below gives the results of these tests. Diameters are in thousandths of 
an inch, and breaking stresses in lbs. per sq. inch. 
Table V.—Strength of Glass Fibres. 
Diameter. 
Breaking Stress. 
Diameter. 
Breaking Stress. 
! 0 - 001 inch. 
lbs. per sq. inch. 
0-001 inch. 
lbs. per sq. inch. 
40-00 
24,900* 
0-95 
117,000 
4-20 
42,300 
0-75 
134,000 
2-78 
50,800 
0-70 
164,000 
2-25 
64,100 
0-60 
185,000 
2-00 
79,600 
0-56 
154,000 
1-85 
88,500 
0-50 
195,000 
1-75 
82,600 
0-38 
232,000 
1 • 40 
85,200 
0-26 
332,000 
1-32 
99,500 
0-165 
498,000 
1-15 
88,700 
0-130 
491,000 
It will be seen that the results are still somewhat irregular. No doubt more precise 
treatment of the fibres would lead to some improvement in this respect, but such 
refinement is scarcely necessary at the present stage. 
The limiting tensile strength of a fibre of the smallest possible (molecular) diameter 
may be obtained approximately from the figures in Table Y. by plotting reciprocals 
of the tensile strength and extrapolating to zero diameter. This maximum strength 
is found to be about 1*6 X 10 6 lbs. per sq. inch, which agrees sufficiently well with 
the rough estimate previously obtained from the cracked plate experiments. 
* From the tensile tests previously described. 
2 c 2 
