MR. C. V. BOYS ON THE RADIO-MICROMETER. 
183 
the extremely feeble force due to the torsion of the fibre is taken into account. I 
have found it absolutely necessary to sink the antimony and bismuth into a little 
well made by drilling a hole, no larger than necessary, in a piece of soft mon which is 
buried in the brasswork. A small lateral hole allows the radiation to fall on the heat¬ 
receiving surface. This effectually screens off the part which produces the greatest 
disturbance (see fig. 7, in which the iron is represented by the darker shading). The 
Fig. 7. 
copper must be exposed to the magnetism; therefore, it must be careftilly examined 
to see that it is neutral, and it must then be kept away from emery or magnetic 
cleaning materials. 
It is finally necessary to show that advantage is gained by employing the antimony- 
bismutb-copper combination, instead of the plain pair of wires used by M. d’Aesonval. 
While antimony-bismuth wires would, on account of their great thermo-electric power, 
be superior to palladium-silver, they would, on account of their magnetic qualities, 
disturb the natural period of the circuit. By the combination which I have employed, 
1 am able to make use of this great thermo-electric power at the same time that the 
magnetic disturbance is avoided. 
Some of the conclusions enunciated m the preliminary note require modification in 
view of the more extended investigation described in this paper. In a note, added 
March 23rd, I had concluded that more than one junction would be advantageous, 
but this is not the case. 
