436 PROFESSOE W. WATSON ON A DETERMINATION OF THE VALUE OF 
into these teeth. Each of the flange castings was bolted to this face-plate and the 
surfaces AB and BC, fig. 2, were turned up. The central hole DE was also turned 
to fit accurately a steel mandrel on which a scpiare threaded sci'ew had been cut. The 
castings were then mounted on the mandrel, and by means of lock nuts were held at 
the right distance apart. A sheet of hard rolled brass, OT centim. thick, the edges 
of which had been cut parallel, was then bent round so as to lie along the surfaces 
AB (fig. 2), and was held in place by screws passing into the castings. The brass 
Fig. 3. 
was ot sufiicient length to overlap for about 30 centims. at the bottom where the hole 
for the passage of the stand had to be cut. The attachment of the cylinder of brass 
to the flanges was rendered additionally secure by tinniiig the surfaces which came 
in contact before putting the parts together, and then, after the insertion of the 
screu's, by heating the coils, solder was run into the joint. The necessaiy holes in 
the cylindrical side of the coils were cut out by means of a milling cutter. By this 
method of construction the two metal flanges, in which the grooves to hold the wire 
were to be turned, were held very rigidly in the same relative positions. 
In order to turn out the grooves the coil was mounted on the mandrel between 
centres and driven by two pins attached to the face-plate. Light cuts were taken, 
particularly at the finish, when the tool being left untouched the coil was reversed 
between the centres so that the depths of the two grooves should be the same. The 
flanges were marked with twelve ecjui-distant lines, parallel to the axis of the coil, to 
act as reference lines when measuring the dimensions of the coil. Two diametral 
lines at right angles were also ruled on each of the flanges near the central hole. 
Plates of mica were then clamped over the central holes by means of brass 
rings, G, fig. 1. The intersection of fine lines ruled on these mica plates, to form a 
continuation of the diametral lines ruled on the flanges, marked the axis of the coils. 
That the rigidity of the coils is quite satisfactory is shown by the measurements 
given in the subsequent pages, Avhich were made eighteen months after the 
