7 
. 
_ON THE ABSOLUTE MEASUREMENT OF HARDNESS. Denk 
the observer can proceed with his work smoothly and without annoying 
accidents; for the weights are added by a method of manipulation 
which is wholly without disturbing influence on the test plates, ete. 
Theoretically, however, the method is not altogether free from objec- 
tions, inasmuch as the impressed surface is continually loaded and un 
loaded, and therefore put through a series of increasing cyclie strains, 
until the final point of rupture is reached. For this reason I made 
special sets of measurements, in which the loads were added gradually 
with every available precaution, of course, but without passing back to 
the pressureless contact at the end of each step. Near the end of the 
operation I carefully charged the seale pan with sand, pouring it in so 
gradually that the increase of the load was practically continuous. 
These experiments led to someinteresting subsidiary facts; but asthey 
did not change the chief issue with which weare now concerned, I need 
only accentuate the remarks already made, that it is never permissible 
to increase the load quickly, not even at the beginning of the work 
when the loads are allsmall. Inall such cases rupture is liable to occur 
prematurely, and the discrepancy is frequently of serious moment. 
In addition to the corresponding values of p and d, P and D, other 
quantities were usually observed, such, for instance, as would be neces- 
sary for corrections, ete., and for the ultimate purposes of this research. 
I always noted the diameter of the Newton rings, as well as the diam- 
eterof the locus of rupture. As to the latter I may here remark (definite 
researc hes will be published elsewhere) that in caseof isotropic media 
it is in fact a circle and concentric with the area of contact; yet it does 
not coincide with limits of this area, but surrounds it in accordance 
with a well-defined law. [For crystals the locus of rupture is not cir- 
cular, but an intermediate figure between a circle and a polygon (hex- 
agon, rhombus, triangle, etc.). 
A few words on the diameter of the spot are in place here because 
of their bearing on the accuracy of measurement. I found by trial that 
the demarcation was sharpest in case of faint illumination, for in this 
case the light was not annoyingly reflected from the upper face of the 
test plate. In general this definition varies in marked degree in differ- 
ent experiments, and even different parts of the edge of a given spot 
are not equally distinct. Usually, however, 0.7 scale part of the microm- 
eter is guaranteed. 
From an economical standpoint it is fortunate that a single plate and 
lens will outlast many experiments, certainly as many as are necessary 
for obtaining a sharp average of results. The lens is not usually af- 
fected, but retains its even surface indefinitely. The plate is large 
enough for upwards of 30 or 40 fields of rupture on each of its sides. It 
is advisable to rule the side of the plate which is not to be used witha 
set of rectangular cross-section lines. Curiously enough, the divers cir- 
cles of rupture do not seem to interfere with each other, even when 
they are nearly contiguous, or when the loci actually intersect. I found 
