.. .[+8r] 
and reaching into the cavity near the centre of the ball. 
The ends of thefe fcrews next the centre are turn’d 
into the form reprefented in the drawing, which, 
prefling with the weight of the ball againfl: the longer 
arms of the levers, caufe the fhorter arms to prefs 
againfl; the end of the brafs bar at 6. Things being 
in this fltuation, let us fuppofe, that the rod of the 
pendulum, and the brafs annexed to it, grow longer 
by heat ; and that the brafs lengthens more than 
the iron of the fame length : then the brafs, by its 
excefs of dilatation, will prefs the fliort ends of the 
levers downwards at 6, and at the fame time necef- 
larily lift up the ball, which refts upon the long ends 
of the fame levers at ff, to any proportion neceflary ; 
And provided the ends of the fcrews do prefs upon 
the levers at a proper diflance from the centres, the 
laid ball will be always kept at the point of fufpen- 
flon, notwithstanding any alteration the rod of the 
pendulum may be liable to from heat or cold. What 
this diftance ought to be, may very nearly be deter- 
mined, if the difference of the expanflon between- 
the brafs and iron bars is known ; for the proportion 
the fhorter arms of the levers ought to bear to the 
longer ones will always be, as the excefs of the ex- 
pansion of the brafs is to the whole expanflon of the 
iron, as may be thus eaiily demonflirated. Fig. 2 . 
Let the line ab, drawn perpendicular to the line ef 
reprefent a bar of iron ; the line c d a bar of brafs, 
the pricked line b g, the expanflon of the brafs 
bar by the fame degree of heat : let the line g i be 
drawn parallel to the line ef then will i h repre- 
fent the difference of the expanflon of the two 
metals: thro’ the points /j,£, draw a right line 
P p p cutting 
