334 MR. W. DUDDELL ON THE RESISTANCE AND 



which it burns, as has been shown by Professor S. P. THOMPSON and others. Putting 

 aside the possible combination of the carbon with the slight trace of impurity, which 

 may account for a. small part of the back E.M.F., the polarisation E.M.F.'s do not 

 seem able to account satisfactorily for the whole back E.M.F. of the arc, without 

 making assumptions which are at present unsupported by any satisfactory evidence. 



Against the back E.M.F. of the arc being due to thermo-electric forces, it is 

 generally urged that these forces are usually reckoned in tenths and hundredths of 

 a volt, and that therefore the order of magnitude of the hack E.M.F., which was 

 then supposed to be about 40 volts, rendered it highly improbable that it was due to 

 thermal causes. 



There is no doubt that the temperature of the positive crater is very high said to 

 be about 3500 C., and it is quite possible that there may exist large temperature 

 gradients near the electrodes, since the true back E.M.F. to be accounted for is only 

 of the order of 17 volts at the positive crater. It is therefore worth while 

 re-considering whether after all the back E.M.F. of the arc may not be due to 

 thermo-electric forces. 



In this connection the experiment of DUBS* is of considerable interest. He took 

 two carbon plates, about 1 millim. apart, and caused a blow-pipe flame to impinge 

 across their edges, and found that a small current could be obtained from one plate to 

 the other through a galvanometer. He considered this result might be analogous 

 to the back E.M.F. of the arc. A similar experiment has also been described 

 by OLIVETTI.! 



The views of the author that impurities, such as the salts of the metals of the 

 alkaline earths, are essential to the existence of the arc, have led him to try a 

 considerable number of experiments on the P.D.'s produced by unequally heating 

 carbon electrodes, either with the addition of such salts to the electrodes, or to the 

 flame used as a source of heat. Most of these experiments were carried out by 

 heating the tips of ordinary arc-lamp carbons held in the hand-feed arc lamp already 

 described, by causing a "blow-through" or a "mixed" oxy-coal-gas flame to impinge 

 upon them. 



The terminals of the arc lamp were connected to a direct-current voltmeter and 

 were not joined in any way to a source of current, so that the P.D.'s observed were 

 not due to leakage from any extraneous source. The choice of the voltmeter presents 

 some difficulty. At first sight, owing to the high resistance of the flame between the 

 two tips of the carbons, it might seem that an electrostatic or a very high resistance 

 voltmeter would be the best. Electrostatic voltmeters were however found unreliable, 

 owing to the friction of the gases in the jet and against the carbons charging them 

 up electrostatically. Two voltmeters were therefore used, the one an ordinary 

 Weston pivot instrument, having a resistance of about 600 ohms, and the second 



* ' Beiblatter,' 1889, vol. 13, p. 197. 



t ' Electrical Review,' 1892, vol. 31, p. 728. 



