84 
PROFESSOR HUGH L. CALLEXDAR OX 
modulus was determined for each section. The tape was wound on each coil in two 
lengths of 19 turns each, starting at opposite ends of a diameter, with two 
thicknesses of paraffined paper between each turn, so that the insulation could be 
tested with absolute certainty at any time. There were 38 complete turns, and 
nearly 6000 centims. of tape, on each coil. The probable error of the measurement 
was less than 1 millim. on the whole length, i.e., less than 1 part in 100,000. The 
coils could not be boiled in paraffin after winding, as this would subject the tape to 
uncertain strains owing to the contraction of the wax. It was found necessary to 
re-wind the coils two or three times with minor improvements before the insulation 
proved to be perfect. Finally, silk ribbon was adopted in place of paper. 
(14.) Distance betiveen Mean Planes of Large Coils. 
The coils were fitted to the frame at a distance between their mean planes equal to 
the mean radius of either coil. For this purpose, the original frame, which was 
insufficiently rigid, was strengthened with heavy rings of brass turned to fit the 
coils, and carefully tested for non-magnetic quality. The distance between the mean 
planes, though four times less important than the mean radius, must be known 
with considerable accuracy. This was effected by making the coils reversible and 
measuring their thickness. Each coil when in position rested against the points ol 
three screws, which were adjusted to the right distance apart, and clamped in position. 
The distances between the points of opposite screws were measured with the callipers, 
and also the thickness of the coils themselves at the points of contact. The coils 
were connected to their respective terminals by eight pairs of flexible conductors, 
twisted together in pairs so as accurately to compensate, and arranged so that the 
coils could be rapidly removed, reversed and replaced at any moment. This appears 
to be the only satisfactory method of determining the distance between the mean 
planes, as the winding of the coils and the distribution of the current in the tape 
cannot be assumed to be perfectly symmetrical with respect to the channels. 
(15.) Area of the Small Coils. 
The most difficult part of the work was the determination of the area of the small 
coils by comparison with the large coils. A thick brass tube, carrying a delicately 
suspended magnet and mirror at its centre, was rigidly fixed to the framework so as 
to be co-axial with the large coils. The small coils could be mounted co-axially 
on this tube at a mean distance of 30 centims. on either side of the centre. Currents 
were passed in opposite directions through the large coils and small coils in 
parallel, and the resistances of the circuits were adjusted until the magneto¬ 
meter showed no deflection. The ratio of the resistances was then immediately 
determined in a manner similar to that employed by Lord Rayleigh in measuring 
