96 Profs. Dewar and Fleming on the 



& 



of a thermo-electric junction to values in the neighbour- 

 hood of — 200° C. by the employment of considerable quan- 

 tities of boiling liquid air and liquid oxygen, we have been 

 able to conduct an extensive investigation on the thermo- 

 electric powers of metals at temperatures lying between this 

 low temperature and that of the boiling-point of water ; 

 and have been thus enabled to continue the thermo-electric 

 diagram down to temperatures not very far removed from 

 the absolute zero. We have accordingly been able to 

 determine the points of electric inversion which lie between 

 0° 0. and — 200° C, and to compare these observed values 

 with those predicted by the extrapolation of the existing 

 thermo-electric diagram of Professor Tait obtained from 

 observations at higher limits of temperature. The chief 

 interest, however, attaching to these investigations was the 

 ability to determine whether, over these w r ide limits of low 

 temperature, the thermo-electric lines of various metals 

 remain straight lines, and whether any of the metals exhibit 

 changes of thermo-electric power at low temperatures which 

 cause their thermo-electric lines to be bent or broken — 

 thus indicating a sudden change in the sign or magnitude of 

 the " Thomson effect" at certain temperatures. The results 

 of these experiments have been set out in a chart of thermo- 

 electromotive forces of the various metals and alloys used in 

 conjunction with pure lead as a thermo-couple, and from which 

 all the required information can be obtained. This chart and 

 the details of observations below provide the data for a full 

 discussion of these questions. 



3. The metals used by us in this investigation were for the 

 most part in the form of wires, and w r ere in many cases taken 

 from the samples of pure metals employed by us in a series of 

 determinations of the specific electrical resistance of metals at 

 low temperature*. Omitting as unnecessary an account of 

 much preliminary investigation, we proceed to describe the 

 final experimental arrangements as follows : — 



A large number of thermo-electric couples were prepared, 

 each consisting of two wires of lead joined by an intermediate 

 wire of another metal M. The lead wire was prepared by 

 pressing pure lead into a wire '0656 centim. in diameter, and 

 the other metallic wires had diameters approximately from 

 about *025 to *15 centim., or else were used in the form of 

 very thin narrow strips. The lead wires had a length of 

 50 centimetres, and the connecting wire of the metal or alloy 



* See Dewar and Fleming, {i The Electrical Resistance of Metals and 

 Alloys at Temperatures approaching the Absolute Zero," Phil. Mag. 

 September 1893, p. 271. 



