180 An Account of Experiments for determining the Length of 
From the above table it seems that 5°51 divisions (or °000236 
of an inch) are to be subtracted from the distance obtained when 
the knife edges are viewed as dark objects on a light ground ; 
and on the contrary, the same quantity to be added when. they 
are seen_as light objects on a dark ground. 
From the few experiments I have made, this quantity appears: 
to be the same, whatever may be the relative illumination of the 
object and its ground, so long as the difference of character is 
preserved. On the cause of this extraordinary fact I can hazard 
no conjecture, and it remains an interesting. subject for future 
investigation. 
Of ihe Expansion of the Pendulum. 
The composition of brass is so various, that probably no two 
specimens possess precisely the same rate of expansion. It be- 
came therefore necessary to determine the expansion of the peu-. 
dulum by direct experiment, instead of adopting the conclusions 
of others, and for this purpose the following method was used, A 
trough of deal was made of a length sufficient to receive the bar 
intended for the pendulum, which was placed edgewise in the. 
middle of the trough, being secured at one end by wedges on. 
both sides. The bar was supported on small pieces of glass , 
tube, serving as rollers to prevent friction, and the trough was of 
the same depth as the width of the bar. 
Two transverse lines were drawn near the extremities of the 
edge of the bar, distant from each other 499 inches, and a third 
line was subsequently drawn one inch beyond. The microscopes, 
were placed over the lines, and left, together witha thermometes;, 
for twenty-four hours previous to the experiment. 
The temperature being then registered, and the microscopes 
having been examined to see that the lines bisected the angles 
formed by the spider’s threads, the trough was filled with hot 
water to the edge of the bar, and two thermometers were placed 
in it, one just beneath the surface of the water, and the other at 
the bottom of the trough. The bar rapidly expanded, and the 
line on it was followed by the micrometer till it became sta- 
tionary. The bisection was then perfected, and the mean of the 
degrees shown by the thermometers registered, together witi. 
the number of revolutions and parts made by the micrometer. 
The whole was now suffered to remain till the temperature had 
become several degrees lower, when the contraction of the bar, 
occasioned by such decrease of temperature, was measured, and 
thus several successive observations were made, which are con- 
tained in the following table, 
Distance 
