94 Proceedings of Poyal Society of Edinburgh. [sess. 
did, or in the opposite one, could a thoroughly constant movement 
of the cylinder be obtained, and the worst of it was that the devia- 
tions from the intended course were without any regularity, and did 
not answer to my expectations. 
I tried afterwards to correct the curves obtained by remeasuring, 
in the same period, the distance of maximum to minimum by 
means of the vertical part of the ocular micrometer referred to. 
But finally I dropped the clockwork altogether, and I use now 
the divisions of the drums, P and Q, exclusively for the measure- 
ment of the abscissae. In order, however, to be able to measure 
ordinates and abscissae of the curve at the same time, I had the 
inner edge of the drum, Q, made so as to be provided with shallow 
grooves made by a file at each of its 360 divisions. In these 
grooves a tooth catches, fastened to a spring, the power of 
which may be regulated by a screw.* By screwing the latter on, 
it is even possible to make the edge of the drum quite free from 
the tooth. If the latter is pressed by the spring against the edge 
it is possible, by turning the drum very carefully with the hand, 
while looking through the microscope, to make it stop at each 
groove made by the file, viz., after of a revolution of the 
cylinder. 
But without the application of a weight the same failure, as 
described above, occurred again, for instead of ten shocks, the drum, 
Q, ought to be turned through a varying number of shocks in order 
to make the drum, P, and the cylinder it carried move through 
exactly % fo ^s circumference. 
After some trials, however, a suitable weight was found, attached 
to a wire running over the wheel, R, which drew the drum, P, round 
in a direction opposite to that in which it was turned by the hand. 
This arrangement caused the teeth of the wheel and pinion to 
cling permanently to each other, so that the movements of the 
cylinder were nearly equal to each other every time that the care- 
fully-turned drum, Q, stopped by the tooth catching the groove. 
The above-described mode of measuring the impressions was used 
exclusively in my later experiments, and it gave reliable results. 
The arrangement by which it was possible to move the cylinder, 
* This arrangement is not to be seen on fig. 2, as it is concealed behind the 
drum, Q, and the wheel, A. 
