1 30 Proceedings of the Royal Society of Victoria. 



there was a sudden swing indicating a change from -0002 to --001, 

 then suddenly back to •016, and steadily on to 034, then 

 a sudden great swing much higher still, but it fell before a 

 reading could be taken. On cooling it fell steadily to zero. 

 When heated again the effect I'ose steadily to -25 volt, and for 

 some time rested steadily between "23 and -2-5. Shifting the 

 flame back and forward increased the effect to -28, and it 

 remained steady at -27 for some time, and on shifting the flame 

 along the whole tube I found two places about 2 mm. apart which 

 gave maximum effects of +-27 and --25 volt. The wire then 

 parted. In repeating the experiment with other samples of the 

 same wire I could never again get the great effects of the second 

 part of the experiment, though "OOl was reached frequently 

 enough. Several times on turning out the gas I noticed that the 

 readings decreased very slowly, and once when the tube could be 

 held comfortably in the fingers -0002 volt was still indicated. I 

 ilipped part of the wire into hot oil at about 250° when -0006 

 volt was given, while at 900° only -001 had been given. 



A tube was tilled with antimony by exhaustion, and on heating 

 it • 27 volt was given almost at once, but not for long. • 1 volt 

 was reached again, but after a time I could get none. Another 

 tube gave "01 volt with the bunsen flame, and -015 with the 

 blowpipe. I made several other efforts to fill tubes with 

 antimony, but failed, as the metal was not continuous through 

 the tube. 



With bismuth the highest effect observed was -06 volt, Init 

 I did not keep at it very long. 



Silver wire heated in a glass tube gave -00007, which, when 

 the flame was removed, rose to '0007. When heated in a clay 

 tube • 0008 was reached after several heatings and coolings. In 

 another piece "001 was reached after a time, but only temporarily. 

 The effect was never steady, though more so than with iron. 

 Another sample of very pure silver gave various effects up to 

 •0001. 



Thallium gave various effects up to -0005 volt, but I did not 

 keep at it long as it attacked the clay, and I did not wish to 

 lose it. 



Magnesium was a difiicult metal to examine. Heated in air 



