1887.] 



On the "Radio-micrometer." 



191 



It is easy to calculate exactly what deflection will be caused by a 

 given rise of temperature in any instrument. Taking quantities, all 

 of which can be easily obtained, namely, antimony and bismuth, each 

 5 X 5 X 0'25 mm., one arch of copper wire one-eighth of a square mm. 

 in section, completing a circuit of one square cm., suspended in 

 a field of 10,000 units by a thread of such a torsion as would 

 give an undamped period of oscillation of 20 seconds, the least move- 

 ment that could be detected would be produced by a rise of tempera- 

 ture of about one ninety-four millionth of a degree centigrade. It 

 seems to be capable of attaining about 100 times the sensitiveness of 

 the bolometer. Further, the electromotive force acting at this tempe- 

 rature would be only about one million millionth of a volt, which 

 is probably smaller than any that can be detected by other 

 means. 



I am now engaged in working out such details as the various 

 maxima. Thus for a certain shape of circuit and number of turns a 

 certain thickness of copper gives the best result ; of various shapes 

 the rectangle only is practically convenient, but taking the besb 

 thickness of wire for each successive length, a certain length is better 

 than any other ; with a particular shape and the best thickness, two 

 turns are better than one or three, but it does not follow that this will 

 be so for all shapes ; no advantage is gained by employing more than 

 one junction. An increased sensibility may be produced by either 

 using a longer torsion thread or a stronger field, a certain relation 

 between these and the resistance will give a maximum quickness. 



The conductivity of aluminium compared to its mass is, as is well 

 known, more than that of any other material, so it would be pre- 

 ferable to copper if it were not for the difficulty of soldering it. 



The instrument may be made with a second active bar, forming 

 the other end of the rectangle, with its own window, and can thus 

 be used differentially, one bar being exposed to the radiation to be 

 measured and the other to any compensator. 



It is interesting to note that the damping action is slightly reduced 

 by the Peltier effect. 



I anticipate some difficulty in using the field produced by an electro- 

 magnet ; for even if the effect of the heating of the coils by the 

 current can be avoided, as I think it may by the use of long iron con- 

 necting-rods passing through water, any variation of current will 

 give rise to motions of the suspended circuit, the nature of which I 

 described in the ' Phil. Mag.,' September, 1884. But after these 

 difficulties are removed, there will remain the diamagnetism of the 

 three metals and the great damping action. With fields of reasonable 

 strength, the diamagnetism may be set against the torsion of the 

 thread, so that but feeble controlling force may be obtained by a 

 thread of moderate length, but for electromagnetic fields I propose to 



