A Nezo Therinoelectric Phenomenon. 127 



steadily to 10 and then jerked to 60, after which it gradually 

 <lecreased to zero. These jerks on the removal of the flame 

 occurred so frequently as to make one of the characteristics of the 

 phenomenon in the case of lead. Sometimes on heating the lead 

 the i-eadings rose gradually from zero, and at others there was no 

 effect for some minutes when a sudden great deflection occurred. 

 Another pipe stem was taken and a lead wire drawn down till it 

 could be pushed along it, and on heating it I at last got -2 volt. 

 At the time this was by far the greatest effect yet observed. On 

 one occasion after heating the tube in the usual way I removed 

 the flame at a time when the reading was steady at zero. The 

 <;.m.f. rose quickly, but not suddenly, to -07 volt, and then 

 decreased. As a type of the general behaviour of lead I will 

 describe the chief movements which took place in forty-five 

 minutes, during which the flame was not disturbed. After a 

 little preliminary heating the needle began to move and indicated 

 •02 volt, then reversed and rose gradually to "17, but decreased to 

 •16, at which it remained steady for some time. It decreased 

 further to *1, but rose to "185 and kept steady at "18 ; decreased 

 to '05, rose to "1, at which it kept for five minutes, then went on 

 to '15, and '18, after which it fell and reversed to -07, but soon 

 came back to -1, and on to -205, fell to -1, rose to -15, fell to -05, 

 but came back to '2, and again fell and reversed to --05 for a 

 few seconds only, after which it rose again to '15 ; reversed again 

 to --08, and --1, and back to -205, remained steady at -203 for 

 half-a-minute. It fell again and reversed to --02, rose to -1, and 

 again reversed to -'IS. The gas was then turned off for the 

 night. There were many smaller motions superposed on the 

 larger ones, but were quite irregular. In spite of the occasional 

 •excursions to the negative side, the deflections were on the one 

 side for the great bulk of the time, - ve and + ve are, of course, 

 quite arbitrary. The motions were all dead beat except some of 

 the negative deflections, which were so quick and so short that 

 they set up oscillations in the needle. When the gas was lighted 

 again next morning, the system not having beer* disturbed, there 

 was no effect after five minutes. The flame was then shifted 

 about, but no effect more than • 0005 volt could be got. I tried 

 several other tubes with lead wires passed through them and 



