228 
MESSRS. R. T. GLAZEBROOK AND J. M. DODDS ON THE 
Let t be the thickness of the wire and silk used, and let A and a be the mean radii 
of the coils. 
Then 2A-| -t — 
26{+ d„+ ... + } + + . . . 
797 
Now we know the diameter of the channel before the first layer was put on and 
also the diameter of the first layer ; they were respectively 
From this we find 
and finally 
49'565 and 49'728 centims. 
£=•0815 centim. 
A=25 - 753 centims. 
The observations on coil B gave the same value for the thickness of the wire and 
covering, and we get for it 
a— 25 - 7CG centims. 
The method here adopted to determine the value of the mean radius allows for the 
fact that in winding one layer may sink somewhat into the one beneath. 
Let the figure (fig. 3) represent a section of the coils by a vertical plane through 
the axis ; let the coils be placed with their lettered sides down as in the figure. This 
we call position 1 throughout. 
In position 2 the lettered side of B was turned uppermost. 
In position 3 the lettered side of A also was uppermost, while in position 4, 
A remained uppermost while B was again inverted. 
Thus if nJ^B mean that the lettered side of B was down, we have 
