C 480 ] 
me to confider, that, as metals differ from each 
other in their denfity, it was highly probable they 
might likewife differ from each other in their ex- 
parffion ; and that this difference of the expanfions of 
two metals might be fo applied, as in a great mea- 
sure to remove thole irregularities in the motion of 
a clock, which arife from the effeCt of heat and coid 
upon the length of a pendulum. With this view, 
not long afterwards I contrived the pendulum now 
defcribed by Fig. 1. 
In which a b reprefents a bar of brafs, made quite 
fad: at the upper part by pins, and held contiguous, 
at leveral equal distances, by the fcrews 1, 2, 3, &c. 
to the rod of the pendulum, which is a bar of iron ; 
and fo far as the brafs bar reaches, is filed of the fame 
fize and fhape, and confequently does not appear in 
the figure ; but a little below the end of the brafs 
bar, the iron is left broader, as atj d d, for the con- 
veniency of fixing the work to it, and is made of a 
fufficient length to pafs quite thro’ the ball of the 
pendulum to c. The holes, 1,2, &c. in the brafs, 
thro’ which the fhanks of the fcrews pafs into the 
iron rod of the pendulum, are filed as in the draw- 
ing, of a length fufficient to fuffer the brafs to con- 
tract and dilate freely by heat and cold under the 
heads of the fcrews. eeee reprefents the ball of the 
pendulum: ff> two ftrong pieces of fteel, or levers, 
whofe inner centres, or pivots, turn in two holes 
drilled in the broad part of the pendulum-rod, and 
their outer ones in a ftrong bridge, or cock, fcrew’d 
upon the fame part of the rod, but omitted in the 
draught ; becaufe, when put on, it covers this mc- 
chanifm. gg t are two fcrews entering at the edge, 
and 
