122 Proceedings of the Royal Society of Victoria. 



After working at the iron for some days I treated a copper 

 wire similarly to see if the effect existed in it too, but could not 

 find a trace of it. The copper wire was a thick one, and this was 

 before I noticed the effect greater in fine wire than coarse. I 

 came to the conclusion that the effect was in some way connected 

 with the magnetic property of iron, and for some time confined 

 my attention to it. After a while, however, I tried a fine brass 

 wire, and instantly found the effect marked, about -0001 volt. 

 As the wire was heated in the naked flame it soon fused. On 

 twisting the ends together and heating the junction -001 volt 

 was indicated. Platinum wires of -8 and -4 mm. diameter gave 

 very small effects, but a very fine wire of '06 mm. diameter gave 

 •0001 volt. Copper wire 1'7 mm. gave no effect, as already stated. 

 •3 mm. gave -00002 volt, and one of -14 mm, -0001 volt. These 

 values were all obtained by heating the wires in a flame, but as 

 all except the platinum fused almost instantly even in a candle 

 flame I had to take steps to protect them. The most obvious 

 plan was to pass them through glass tubes, but at a red heat 

 there seemed to be chemical action between some of the metals 

 and the glass, so the glass tubes were abandoned for clay tobacco 

 pipe stems. Even the finest wires could be heated for some time 

 in these without burning through or fusing. 



Gold wire, when heated, presented some interesting peculiarities. 

 The first tried was an alloy of gold and silver, 62 per cent, 

 gold (fifteen carats). It was somewhat fine wire, • 26 mm. diameter. 

 I found the eflect well marked, though at first not so great as in 

 iron, but more steady. The effect was not constant, but the 

 changes took place very slowly, so that the galvanometer needle 

 moved dead beat. Repeated heating and cooling the same part 

 greatly increased the effect, this was not noticed in iron. 

 Frequently on cooling the tube I noticed an extraordinary effect. 

 On turning off the gas there were almost immediate, and 

 apparently instantaneous, rises, sometimes of fifty per cent., some- 

 times as much as one hundred per cent., though only temporary ; 

 one of these sudden rises reached -01 volt. While repeating 

 the experiment the wire fused, and I had to take a fresh piece. 

 On several occasions I found that by shifting the flame back and 

 forward over an inch or two of the wire that there were points 

 which gave a maximum effect, intermediate points giving little or 



