AND A SELECTION EEOM THE OBSERVATIONS MADE WITH THEM. 
697 
abraded matter would not be equally distributed, and, collecting in excess in some place, 
unequal action would be set up before the pitch had time to yield. If the excess was 
not at the centre, the depression would assume the character of a ring. The pitch at 
length yielding, the ring would not increase, but it would continue, and, a similar cause 
arising in another part of the polisher, a second ring would be formed, and so on. I have 
seen a surface of an annular character all over, the breadth of the rings depending on 
the adjustments of the eccentrics. Why the depressions once formed continue with so 
much persistency is evidently owing to the yielding character of the pitch, which, when 
the depression is of large area, becomes protuberant, precisely as it does where it overhangs 
the speculum, and so the cutting action is to some extent continued. An annular 
surface is produced by grinding, under similar cu’cumstances, but the rings change 
their places frequently. The annular surface is always best-marked when the action of 
the second eccentric is suspended completely. To see the annular surface, the speculum 
must be slightly polished by rubbing it all over for a few minutes with a small lap 
covered with soft pitch and rouge. I need, perhaps, hardly add that the character of 
these surfaces can only be seen when they are examined by the light reflected from the 
watch-dial, in the way described in the ‘Transactions’ for 1840. As the throw of the 
eccentrics is increased the rings gradually disappear ; and when they reach the proper 
positions the surface becomes quite uniform. We have often found it very useful, when 
the figure of the polisher was not satisfactory, to throw another movement into gear con- 
nected with the guide D, by which an occasional stroke was given of increased length : 
the cause of unequal action is thus immediately removed if it does not arise from some 
defect in the construction of the polisher, such as the contact of some unyielding sub- 
stance with the iron. The experiments I have just referred to were of a very early date, 
but they were numerous and made with great care ; I have therefore not thought it 
necessary to repeat them. 
In the first polishing-machine we made, the polisher was connected with the eccentric 
B by means of a rigid bar passing through the guide D, the guide being furnished with 
an adjustment at right angles to the line joining the centres of the speculum and 
eccentric. The guide was equidistant from the centres of the eccentric and polisher, 
and the path of a point in the polisher was similar to that of the crank-pin of the 
eccentric. We found, however, that when the movements were very small the surfaces 
both in grinding and polishing became somewhat annular, and when the movements 
were large the figure was spoiled. We therefore substituted a jointed rod for the 
rigid bar, and added the second eccentric. From time to time we have returned to the 
rigid bar, tempted by its obvious advantages, and hoping in some degree to free it from 
its defects. It is an advantage that with it the movement of a point in the polisher is 
as the circumference of a circle, while with the jointed bar under similar circumstances 
it is as the diameter. In the one case there will be more than three times the amount 
of motion there will be in the other, and the polisher will overhang the speculum but 
for a moment at each stroke, instead of dwelling for a much longer time twice during 
5 B 2 
