72 
PROFESSOR HUGH L. CALLENDAR ON 
greater than 10° C.G.S. units. This divergence might be explained as due to 
imperfect insulation if the precautions taken had not been so great, but it more 
probably represents the order of accuracy at present attainable in the absolute 
measurement of resistance by this or any other method. It is hardly possible 
that it could be entirely accounted for by errors of measurement of the diameter of 
the coil and the disc, although a discrepancy of 1 in 10,000 was actually found 
between the calculated and measured diameters of the coil (‘ B.A. Rep.,’ 1897, 
p. 217), which points to some uncertainty in this direction. 
At the time when I first examined the coil of the Lorenz apparatus, some time 
after its arrival in Montreal, while it was being set up, it seemed to me that the wire 
had worked a little loose on the marble, owing to some effect of the drying of the 
insulating tape and varnish, or to straining of the soft insulated wire due to 
contraction on exposure to cold. This would necessarily occur owing to the great 
difference in the coefficients of expansion of the wire and the marble, and the very 
small limits of elasticity of the soft annealed wire. Professor V. Jones himself, 
with whom I discussed the question in September, 1897, shortly after the arrival of 
the apparatus in Montreal, considered that the diameter could not be satisfactorily 
measured with a silk-covered wire, and strongly recommended the re-winding of the 
coil with bare wire. Accordingly, I procured for this purpose a sample of highly 
elastic silicium-bronze wire of high conductivity. I satisfied myself that the limits 
of elasticity of this wire would be ample, if it were wound on under suitable tension 
at a suitable temperature, to keep it perfectly tight on the marble cylinder for any 
range of temperature to which it was likely to be exposed. The tightness of the 
wire in practice is most important, from the point of view of insulation as well as 
from that of accurate measurement of the diameter. If the diameter of the wire is 
nearly equal (as it must necessarily he) to the pitch of the screw thread, a very 
slight defect in tightness or straightness will produce a short circuit. It is quite a 
difficult matter to wind a perfect coil of 200 turns of this size, unless the wire is 
highly elastic and quite free from kinks. 
Owing to the great importance of securing perfect insulation, I proposed to adopt 
the method which I had already put in practice in the case of the electro- 
dynamometer, namely, to wind the coil in a double screw thread with two separate 
wires, in order to have a check on the perfection of the insulation, which could 
be applied at any time after the coil had been wound, or at any moment during 
the actual experiments. 
In consequence of the delay caused by the failure of insulation of the first coil, 
rhe apparatus did not arrive in Montreal until the beginning of September, 1897. 
Some time was occupied in the course of the winter in building a suitable pier, and 
setting up the apparatus. But when I was about to commence observations, I 
received news of my appointment to the Chair of Physics at University College, 
London, which made it necessary for me to abandon the work. 
