804 
MR. H. TOMLINSON ON THE INFLUENCE OF STRESS AND 
must within these limits completely compensate one another. In fact, however, there 
results a diminution of the amplitude of oscillation; hence the loss of energy 
therein can only correspond to the work which is expended in the alteration of the 
positions of equilibrium, or the rotation of the molecules from -\-(3 to —/3 which 
determine it. The diminution of amplitude may therefore be taken as a measure of 
this work. Since the shifting of the permanent torsion-position is, within certain 
narrow limits, proportional to the temporary torsion, and therefore the rotation of the 
molecules therewith is likewise approximately so, the amplitudes must, within those 
limits, diminish according to the law of a geometrical series.” 
Again, Professor Wiedemann agrees with Sir W. Thomson that, whilst an increase of 
load at first increases the loss of energy, this loss, under repeated oscillations, becomes 
less and less, and finally approaches in magnitude that obtained with the lighter load. 
Wiedemann’s explanation of the matter is that, just as in his own statical experi¬ 
ments the torsional displacements were found to be initially greater with a heavier 
load than with a lighter one, so also must the decrements attendant upon oscillations 
be greater in the former case, but, like those, gradually approach the magnitude which 
is observed with the lighter loading. 
Briefly expressed, then, the views of Sir W. Thomson and Professor Wiedemann 
are as follows :—The former, whilst acknowledging that his experiments do not show 
that the proportionate loss of energy of a wire when vibrating torsionally is inversely 
as the period of vibration, which should be the case if the internal friction of a solid 
resemble that of a fluid, still employs the term viscosity with reference to metals, and 
appears to think that the internal friction of a metal partakes of the nature of the 
viscosity of a fluid ; the latter is of opinion that the loss of energy is mainly due to a 
shifting from side to side of the positions of stable equilibrium of the molecules 
caused by the torsional vibrations of the wire. 
I now proceed to give an account of my own researches relating to the internal 
friction of metals. 
Description of Apparatus and Mode of Experimenting. 
The wire under examination was firmly clamped at its upper extremity to a stout 
brass block, which was itself clamped to the end of a strong iron bracket projecting 
from the outside of the wall at the back of my house, the height of the bracket above 
the ground being 685*5 ceutims. A wooden box A A, nearly 600 centims. in length 
and 12 centims. square inside, protected the vertically suspended wire from currents 
of air, and passed through the roof of a small wooden house. The box was secured 
at its upper extremity to the iron bracket, and rested with its base B B upon another 
box SS (fig. l), which measured about 40 centims. in length and breadth and about 
86 centims. in height. The box S S had in front a glass window W W and a door, 
and at each of the two sides a door.* All the doors were lined with caoutchouc, so 
* These doors are not shown in the figure. 
