on the Thermo-electric Qualities of Metals. 325 



motive force at the terminals of the galvanometer as now used is 

 0'135 mikrovolt per division. 



7. A considerable lag was found in the thermometer readings, and 

 the following method was adopted to get rid of this effect. The hot 

 junction was heated very slowly through a small range (5 C. or 10 C.) 

 and then the Bunsen burner was drawn slightly aside so as to give 

 approximately as much heat to the vessel which contained the glycerine 

 as it lost by radiation. The thermometer, T, and the spot of light 

 on the scale were simultaneously observed, and when both were seen 

 to be steady, the readings were noted. The circuit was then imme- 

 diately broken and readings taken of the galvanometer zero* and the 

 thermometer T'. 



Very often the glycerine was allowed to cool slowly, and by means 

 of a small Bunsen flame the temperature was kept steady for a short 

 time and readings taken. If the same precautions were taken as are 

 described in the previous paragraph, the same deflections from zero 

 were got for the same difference of temperature between the hot and 

 cold junctions. 



8. The metals so far tried are : 



(1) Copper wire from Messrs. Johnson and Matthey. This was 



pure electrotype copper wire with no impurity detected except 

 an unweighable trace of iron. 



(2) Copper wire, ordinary commercial, from Messrs. Johnson and 



Matthey. This was analysed! in the chemical laboratory of 

 the University, and was found to contain : 



* It was found necessary to take the zero immediately after each reading as the 

 zero was by no means constant. It was thought at first that the change of zero 

 was mainly due to the suspending fibre of the mirror in the galvanometer, but a 

 new plug and fibre did not lessen the variations of the zero during an experiment. 

 It is most likely due to the general laboratory experiments going on simultaneously, 

 which involve the moving of apparatus and the walking about of students with 

 knives and keys in their pockets in the near vicinity of the galvanometer. For 

 example, my own pocket knife at the distance of the scale from the galvanometer 

 (a metre) gives a deflection of 10 scale divisions. The galvanometer is at a distance 

 of 11'74 metres from the dynamo used in the electric light installation of the 

 Physical Laboratory, the two being separated by a stone wall. The stopping or the 

 starting of the dynamo altered the metallic zero of the galvanometer by 100 

 divisions. The constant of the galvanometer was tested both when the dynamo 

 was running and when the dynamo was not running. Practically it was the same 

 on both occasions. Nearly all the experiments were done when the dynamo was 

 running. 



f All the chemical analyses stated in this paper were given to me by Mr. 

 Anderson, of the Chemical Laboratory of this University. 



2 c 2 



