270 
NATURE 
curve passing through the zero point of stress and strain, just as 
do the elastic ones. It appears as if the #on-elastic, the flowing 
or plastic state, were the rea? state of the material, the edastic 
condition being something consequent on the treatment which 
S.C:CROWN IRON BAR 
NO. 8391. 
BREAKING LOAD 20.87 TONS PER SQ. INCH 
EXTENSION IN IO INS 24.8 PER CENT. 
REDUCTION OF AREA AT FRACTURE 50.5 PER CENT. 
EXTENSIONS 
i 
0) Tons 
5 a 15 per sq. in, 
Fic. 3. 
the materiel had undergone. i 
limit of elasticity 
I have no distinct evidence connecting the ratio . 
maximum load 
with the amount of previous work done on a material in manu- | 
facture. 
3,0 INCHE 
== =4 
Rasa 5= 
EXTENSIONS 
o 20,000 49,000 {p00 3,000 20,060 
40,000 LBS. 
PER SQ.INCH 
I should like to add that the credit of scheming and working 
out the very ingenious apparatus by which the autographic 
diagrams figured above were recorded belongs almost entirely to 
my friend Mr. A. G. Ashcroft, who has been working at the 
matter for me for some time. ALEX. B. W. KENNEDY 
Spectra Produced in Glass by Scratching 
_A FEW weeks ago, while examining under the microscope a 
piece of glass on which a coarse scratch had been made by a 
file, in order to serve as a focussing mark in the determination 
I am bound to say, however, that | 
faintly coloured spectra running along both sides of the scratch. 
As I can find no account of such an appearance, and an exami- 
nation of it seemed to throw a little light on the effects of a 
combined tangential stress and pressure on a brittle medium such 
as glass, I thought a short description of the phenomena might 
be interesting to the readers of NATURE. 
The spectra (Figs. 1 and 2) run for the most part approxim- 
ately parallel to the scratch, but those near the scratch are very 
much curved, the concavity being inwards, and often appear to 
commence and terminate at irregularities in the scratch. If the 
glass be left to itself after scratching, the spectra sometimes 
remain stationary, but not unfrequently spread outwards from 
the scratch ; this process I have watched in three instances. In 
any case, sooner or later, the glass splits internally along the 
edge of the outermost spectrum, and sometimes along the 
others also. I was fortunate enough to watch this splitting in 
one instance: immediately before it took place the glass gave 
signs of great activity, the spectra waving about in the field of 
al 
Fics. 1 and 2.—The shaded part in all the figures represents the scratch. 
view about three times in as many seconds, oscillating between 
two extreme positions (@ and 4 in Fig. 3), and finally coming to 
rest in the position 4, while the split developed with great rapidity 
from above downwards in the field of view. After this splitting 
has taken place, the spreading of the spectra ceases, and they 
generally, though not invariably, remain apparently unaltered. 
The time which elapses between the infliction of the scratch and 
the development of the split varies from a few minutes to several 
days or weeks. 
The appearance is not shown by all scratches, but only by 
such as have produced considerable disturbance in the glass : 
thus they must be fairly deep and must produce some slight 
amount of splintering. 
Next, as to the explanation of the phenomenon. Diffraction 
from the scratch is negatived by the great distance from the 
scratch to the spectra, and still more by the fact that they are 
farther apart the greater their distance from the slit; this im- 
portant point was determined by careful measurement with a 
of refractive index, I noticed a number of narrow, somewhat | micrometer, using sodium light. They are clearly not due to 
[Fuly 23, 1885 
