1911-12.] On the Torsional Oscillations of Magnesium Wire. 247 
XVIII. — On the Torsional Oscillations of Magnesium Wire. 
By G. P. Seaman. Communicated by Professor W. Peddie. 
(MS. received Feb. 5, 1912. Read Feb. 5, 1912.) 
It has been shown by numerous experiments (see Proc. Roy. Soc. Edin., 
1911) that the equation y n (x-\-a) = b holds very accurately for wires of 
different materials, where y = range of oscillation, x = number of oscillations, 
n, a , and b, constants depending on the metal under the given conditions. 
In the following experiments a magnesium wire was subjected to various 
conditions, e.g. extensional strain, torsion, change of temperature, etc., and 
in each case the above law was tested. The method of procedure was as 
follows A magnesium wire of known dimensions and under ordinary 
physical conditions was fixed at its upper end to a clamp, and to its lower 
end was attached centrally a long horizontal heavy cylinder. To the 
cylinder a circular scale divided into millimetres was attached concentrically 
with the wire. When the upper end of the wire was turned round the 
cylinder was made to rotate round the wire as an axis, and the reading on 
the scale, when the maximum angular deflection was reached, was taken 
by means of a telescope. Similarly, consecutive maxima were read until 
the oscillations became small, when the zero was taken, and thus successive 
values of y were found. 
The experiments were directed towards finding the effects, on the con- 
stants n, a, and b, of the subjection of the wire to (1) original conditions in 
the unheated state, (2) repeated rotational strain in the unheated state, 
(3) repeated extensional strain in the unheated condition, (4) a temperature 
of 100° C., (5) a temperature of 250° C., (6) a temperature of 350° C., (7) a 
temperature of 450° C., (8) a temperature of 500° C. The same wire was 
then immersed in liquid air for some time, but this produced no evident 
effect upon the constants. Since we have nlogy -\-\og(x-r a)=>logb, if 
log y be plotted against log (x -f a) the points will, if the formula hold, lie 
on a straight line provided the proper value of a be chosen; and the 
tangent of the angle which this line makes with the log y axis will give n, 
whence b can be calculated. In each case a duplicate set of observations 
was taken so as to confirm the results. 
Numerical Results. 
(i.) In the case of magnesium wire under normal physical conditions, it 
was found, on plotting logy against log (x + a), that the whole graph did 
