280 ON THE RIGIDITY 
nearly as possible,in both. In the static determinations of this, 
and of all subsequent series, Mallock’s mode of applying the 
twisting foree was used. The kinetic observations were made 
with different amplitudes of angle of oscillation, and the static 
observations with differeat angles of torsion. The general 
procedure was as in the former case. Table II contains the 
results. 
In none of the columns of rigidity values of this table 
do the values found indicate any simple law of variation with 
tension. They do not even increase or decrease continuously as 
tension increases, but appear to oscillate between increment and 
decrement, and by amounts which are not accounted for merely 
by errors of observation. The variations cannot be accounted 
for even by errors of method, because in general both methods 
give similar variations. They may, perhaps, be partially at 
least, accounted for by defective procedure. Nevertheless, two 
conclusions may be drawn :—(1) The smaller the angle of 
torsion in the static determinations and the angle of oscillation 
in the kinetic determinations, the greater is the value of the 
rigidity obtained. Mallock drew the same conclusion as to 
kinetic rigidities from his observations. (2) The kinetic 
determinations show a point of minimum rigidity as tension 
increases; but the static determinations are not sufficiently 
exact to be decisive as to whether or not the existence of this 
point is independent of the method. Thus the kinetic observa- 
tions bear out Mr. Macdonald’s result that the kinetic rigidity 
exhibits the minimum point in the case of a cord previously 
unstretched. 
_ The cord used in the last experiment being now in a state of 
tension, was experimented on in a reverse manner. It was left 
a day under the full load, when its rigidity was determined. 
Then one of the weights was taken off, and it was again left for 
a day under the diminished load, and its rigidity determined ; 
and so on, 
