244 
PROFESSOK J. A. EWIXG AND ME. J. C. ME HEMIFEEY OX THE 
J3efore the experiments Avere made it had been conjectured that the de.striictive 
effect of repeated alternations of stress might he ascribed to a loosening of the inter- 
crystalline cement rather than to damage of indi^ddual crY.stals. PreAnoiis exjDeri- 
ments had shown that in fracture by ordinary 2 :)rogressively augmented strain the 
material gives way, in general, not at boundaries, but through the crystals them¬ 
selves,'^ but it seemed po.s.sible that the effect of repeated straining might be different 
in this re.spect. By way of testing the point the experiment was made of .subjecting 
an unetched .specimen to majiy rever.sals of stress, in order to see wliether the inter- 
crystalline boundaries became apparent as they would do if yielding took place 
between each crystal and its neio-hbours. Nothing of the kind was seen, though the 
l^oundaries in some instances could l^e traced through the development in different 
directions of .slip-bands over individual neighbouring crystals. And later experi¬ 
ments, v.'hich will now be described, demonstrated that the mischief which is done by 
repeated straining occurs in cpiite a. different way. 
In experiments made with stre.sses ranging from 14 down to 9 tons per sq. inch it 
was found that fracture ultimately resulted in all cases. The course of the break¬ 
down was as follows ;—Tlie first examination, made after a few reversals of the stress, 
.showed slips-lines on some of the cin^stals, on many of them if the stress was 
comparatively great, but on a few only if the stress did not much exceed the lower 
limit named above (of 9 tons per sq. inch). At this early stage the slip-lines were 
(piite similar in apiiearance to those which are seen when a simple tensile stress 
exceeding the elastic limit is applied. Viewed under vertical illumination they 
appeared as fine dark lines. After more reversals of stress additional .slip-lines 
appeared, which had not been visible in the first instance, but the most conspicuous 
feature was that those which were visible before became far more distinct and showed 
a tendency to liroaden. After many rever.sals they changed into comparatively wide 
bands vrith rather hazily defined edges, losing entirely the fine and sharp character 
v.hich slip-lines })resent when they first appear. As tlie number of reversals 
increased this process of broadening continued, and some parts of the surface became 
almost covered with dark markings made up of groups of broadened lines. Vdlien 
this stage was reached it was found that some of the crystals had cracked. The 
ci'acks occurred along Ijroadened slip-bands ; in some instances they were first seen on 
a single crystal, but soon they joined up from crystal to crystal, until finally a long 
continuous crack was developed across the surfree of the .sjiecimen. AYheii this 
happened a few more reversals brought about fracture. 
In this description we have provisionally named a lower limit of 9 tons jjer sq. inch, 
but the experiments give grounds for belie^■ing tliat an even smaller stress will 
produce fracture in a similar manner if the process of reversing the stress is continued 
sufiicientlv long. There is clear evidence that with 9 tons per sq. inch fracture 
results. But we have also observed tliat with 8, and even 7, tons per sq. inch slip- 
* Ewixg and Eosexhaix, he. cif., p. 372. 
