t 215 ] 
XVI. — On the elasticity of threads of glass , with some of the most useful appli- 
cations of this property to torsion balances. By William Ritchie, A.M. 
F.R.S., Rector of the Royal Academy of Tain . 
Read March 18, 1830. 
i. From facts connected with crystallization and elasticity, it seems ex- 
tremely probable, that the atoms of matter do not attract each other indiffe- 
rently on all sides. There appears to be peculiar points on their surfaces 
which have a more powerful attraction for each other, than for other points on 
the same molecule. This property is not peculiar to the atoms of ponderable 
matter, but seems also to belong to those of light and heat. It is as impossible 
to prove directly the existence of this property, as it is to prove the existence 
of atoms themselves ; but on account of the satisfactory manner in which it 
enables us to explain the'phsenomena of crystallization and elasticity, it is now 
generally adopted. 
2. If the atoms of solid bodies be slightly displaced by any mechanical 
means, they will endeavour to return to their former state of aggregation when 
the disturbing cause is removed. This property belongs in very different de- 
grees to different substances. In lead it scarcely exists, and but slightly so in 
soft copper. In brass, iron and silver, especially when drawn into wires, it 
exists in a considerable degree. But all these substances have limits beyond 
which the property does not hold. If for example an iron wire be twisted 
several times, it will not return exactly to its former state, but remain partially 
twisted. In threads of glass, on the contrary, there seems to be 
no limit to this property, whilst the thread remains entire. Let 
a fine glass thread be suspended from a moveable index, and 
let a light horizontal needle of wood or straw be fixed to its 
lower extremity as in the annexed figure. If the thread be now 
twisted by means of the handle H, whilst the needle N is pre- 
vented turning round, and then allowed to untwist itself, the 
needle will return exactly to its former position after it has ceased to oscillate 
ST 
