an improved Thermometer. 8 i 
and kk into 1 4 equal parts, beginning at 0, the point where il 
cuts the line dh, continuing afterward fix diviiions now on 
that line below kk, making in all 20 equal divifions. If now 
lines be drawn through each of the dividing points, from 0 to 
20 to the line cc, at right angles with the fame, they will give 
on the line cc the true thermometrical fcale to every 5 0 from 2 
to 102, properly adjufted to the inequality of the tube*, which 
in this cafe is nearly of the fame diameter at each end, but 
fmaller towards the middle. Tubes may indeed be found of 
fome conliderable length witli lefs inequality than what this 
fcale exhibits; but the error is here enlarged, to render the 
method of correcting it more confpicuous. 
* Experimentally to prove this method I have made mercurial thermometers, 
whole lcales from the freezing point to that of boiling heat were nearly three 
feet, and though the inequalities of the tubes were very confiderable, varying in 
contrary diredtions to each other; yet when they were placed on the fame frame, 
they perfectly agreed in a motion of the mercury in every part of their fcales. 
Vor.. LXX 1 I. 
M 
