220 MR. RITCHIE ON THE ELASTICITY OF THREADS OF GLASS, 
to be attached, whilst its other extremity is to be fixed to a spiral spring formed 
of fine brass wire, in order to keep the glass thread properly stretched. The 
knife edge is made to rest on two small portions of thermometer tube, placed 
parallel to each other on the top of an upright support. Portions of a still 
smaller steel blade are to be fixed, nearly in the same straight line with the 
edge of the fulcrum and at the ends of the beam, for suspending the scales. 
In one end of the beam a fine sewing-needle is to be fixed, for the purpose of 
pointing out, on a divided scale, the position of the beam when nearly hori- 
zontal. A similar needle is to be fixed in the cylindrical key, for pointing out 
the number of degrees of torsion which the thread has suffered. 
The annexed figure represents a perspective view of the instrument seen 
obliquely, in which b is a beam, g the glass thread, k the knife edge fulcrum, 
t t' the thermometer tubes for supporting the beam, s the spiral spring with its 
attached silk thread, c the divided circle, and a the torsion key. 
Having thus described the balance itself, we shall now explain the manner 
of employing it; first, for the determination of very small weights, and secondly, 
for heavier ones. 
1st. To determine the weight of a small quantity of matter by employing 
only one weight, suppose a grain. — Twist the glass thread by turning the tor- 
sion key through two or three circumferences, according to the degree of torsion 
which it will bear without the risk of breaking. Put brass filings or other 
convenient counterpoise, into one of the scales till the beam be brought to a 
position nearly horizontal, the index of the torsion key pointing to zero on the 
divided circle. Place the body to be weighed into the scale, which is to be 
raised by untwisting the glass thread. Turn the torsion key till the elastic 
force of the thread raise the weight, and carefully observe the point on the 
scale at which the needle, in the end of the beam, becomes stationary. Note 
