140 Dr. J. Erskine-Murray. 



70. As a rule the upper plate was heated, in order that the 

 lower plate might not be affected by draughts of hot air, as would 

 have been the case if the lower had been hot and the- up per cold. 

 Sometimes, during the time of cooling, the upper part of the 

 apparatus, including of course the upper plate, was removed after 

 each observation and was replaced only the moment before the next. 

 In other experiments the upper plate was merely drawn up as far as 

 possible (about 10 cm.) to prevent its warming the lower one. The 

 temperature of the lower plate was also observed, but as a rule it 

 varied only a very few degrees. 



71. At first I used to apply a hot soldering bolt to the back of 

 the plate in order to heat it, but latterly I heated two or three small 

 blocks of tinned copper and placed them on the back of the 

 upper plate. By this second method it was possible to observe the 

 variation of the potential during the rise as well as the fall of 

 temperature. 



72. In attempting to determine the temperature- coefficient of 

 copper we were long baffled by curious anomalies. Sometimes the plate 

 was positive when hot, other times negative, and occasionally it did 

 not vary at all. The clue to this was found in observing that during 

 one experiment while the copper was cooling it was at first positive, 

 then negative, and then it gradually became positive again, though 

 never quite reaching its original value. 



Now it had been found that copper oxide is negative to copper, 

 and that it became temporarily more negative when hot ; hence it 

 was guessed that the successively positive and negative variation 

 must be due to hot clean copper being positive to cold copper, but 

 that it had finally become oxidized and therefore negative whether 

 hot or cold, the small permanent change being due to the thin 

 coating of oxide formed. 



73. The copper was heated much more gently next time, and 

 gave the expected result that clean copper becomes rapidly more 

 positive as its temperature rises, and that, on cooling, its potential 

 returns to its original value unless the temperature has exceeded a 

 certain limit. If this limit has been exceeded its potential rapidly 

 becomes negative and does not return to its original value. 



74. At ordinary atmospheric temperatures the surface of clean 

 copper remains for a long time almost unaltered either visibly or 

 electrically (see 31), and the film which ultimately forms on the 

 surface cannot be pure copper oxide, because the potential of tarnished 

 copper is higher than that of copper oxide obtained by neating in 

 air. If, however, the temperature of the copper is raised to about 

 80 C. it immediately begins to oxidize, though heating to a tempera- 

 ture below this limit does not rapidly produce any permanent change. 

 Thus there is, as it were, an ignition point for copper and oxygen in 



