METEORIC AND ARTIFICIAL NICKEL-IRON ALLOYS. 
35 
each edge of the two cylinders. The distance between consecutive notches in each 
edge was proportional to the circumference of each cylinder, and their number was 
equal to the number of turns of wire it was proposed to wind round the ring. Two 
annuli of mica were next cut of such size as to fit closely between the mica cylinders. 
The cylinders and annuli were then fitted over the ring and were kept in position 
finally by the windings of platinum wire which enveloped them. The second and 
third coils were wound in the same way as the first. The terminals of each coil were 
brought out through holes pierced in one of the annuli serving for the winding and 
insulation of the coil wound over it. In this way it was found to be easy, after 
practice, to wind the permeability coils and the heating coil fairly uniformly and with 
very perfect insulation from one another—nothing but mica being used in the 
insulation. In the earlier experiments, layers of thin asbestos ribbon were wound 
between each mica insulation, partly because under these circumstances it was easier 
to keep the mica of the next coil in position during the winding of that coil, and 
partly to diminish the possibility of short-circuiting between consecutive coils. The 
winding of the coils being completed, the ring was placed upon a sheet of mica 
through which holes for the passage of the terminals of the coils were pierced. This 
sheet of mica was then laid on a small wire tripod, which rested on a block of hard 
wood about 5 centims. thick. This block of wood was also supported on a tripod, 
and thick insulated copper leads projected through it vertically below the ring. The 
outer ends of these leads were provided with binding screws, while the inner end of 
each was connected to a coil terminal. (There were altogether eight of these leads, 
six for the coil terminals and two for the short platinum loop used for compensation 
when determining the temperature of the ring secondary by measuring its resistance.) 
A large inverted beaker was placed on the top of the block of wood, and rested in 
a deep circular groove about half-filled with mercury. By this means the ring, which 
was nearly at the centre of the beaker, was kept out of contact with the outside air. 
There were, however, two glass tubes, leading from outside into the beaker through 
the wooden base, by means of which the air within the beaker could be withdrawn 
and replaced by dry air or other gas. The beaker was kept in a fixed position with 
respect to the baseboard by an asbestos disc placed above it and attached by wires 
to the board. 
A rectangular cover of asbestos board was fitted on the top of the wooden base¬ 
board, and by this means a layer of air about 5 centims. thick was enclosed between 
the beaker and the outer atmosphere. Inmost of the experiments a smaller asbestos 
box was placed inside the beaker and enveloped the ring. 
§ 4. In this apparatus, with a given current in the heating coil, the temperature 
usually became practically constant in less than an hour, but in most cases the 
permeability measurements were not taken until about two hours after the heating 
was begun. 
The relation between the steady temperature in the apparatus and the current 
F 2 
