( 97 ) 
Glafs, or diflurbing the Box. By this means both the 
fharpnefs of the Point, and polifh of the Cap were 
better preferved from injury, when there was occafi- 
on to move the Box. A lixiall piece of Brafs was 
made to Bide upon that End of the Needle which 
pointed to the South, for readily bringing it to an ho- 
rizontal Pofttion ; for according tothe different ftrength 
•of the Touch, the North End of the Needle will dip 
more or lefs. The bottom Plate of Brafs was a little 
broader and longer than the Box, and its Edges made 
Lines exactly parallel to the middle Line of the Divi- 
fions ; and for the greater fecurity of placing the Box 
in a right Situation, there was a Brafs Ruler of thirty 
Inches long, having its Edges even and parallel, ex- 
cept part of that Edge which was applied to the Side 
of the Box which was a little filed away on the Mid- 
dle, that the Side of the Box near its Ends only might 
touch the Ruler. By this Contrivance the two Points 
of Contact were as far afunder as the length of the 
Box would admit of, and the other Edge of the Ruler 
making a longer Line than the Side of the Box, af^ 
forded a better Direction forgiving it the fame Situ- 
ation. 
For determining the Quantity of the Variation, I 
got a Meridian Line ftretched upon the top of the 
Houfc, between the Rails of the Leads, which were 
above fifteen Foot alunder, and the Line was a little 
more than thirty-nine Inches above the Leads. As 
this Line was faftned to two Pieces of Brafs that were 
fixed in the Rails, and was above fifteen Foot long, 
no fenfible Error could arife in putting it up at any 
Time. The Compafs-Box was placed upon a Wooden 
Stool, w T ith three Feet, that had nothing of Iron about 
it, and its Top fet level by a Plumb-R^ile. But find- 
Q z ing 
