Oady and Arnold — Electric Arc. 401 



if it were shown that at the temperature in question the anode 

 emits not negative but positive ions.* 



§28. A test for ionization in the neighborhood of the iron 

 arc was made after the method of Child. f ISTo perceptible 

 excess of positive over negative ionization could be detected, 

 either on the first or second stage. In this connection it would 

 be of interest to measure the rate of emission of charged 

 particles from iron oxide at different temperatures. 



As for the absolute value of the anode drop on the first stage, 

 this cannot be assumed to be the same as on the glow discharge, 

 on account of differences in temperature, current and compo- 

 sition of the gas in the neighborhood of the anode, as well as 

 the presence of ultra-violet light from the cathode. Observations 

 made by means of a platinum exploring electrode in an arc 

 with a silver anode indicated that the anode drop is less on the 

 first stage of the arc than on the glow discharge. This would 

 be expected, in the light of Skinner's observations on the con- 

 ditions that control the drop at the anode4 In our experiments 

 it was not possible to obtain voltage readings of more than 

 relative value. 



§29. Melting Point and Heat Conductivity. — Much weight 

 has been laid by Grandquist and others on the importance of 

 the parts played in the mechanism of the electric arc by the 

 heat conductivity and melting point of the electrodes. It is 

 of interest to examine the connection between these quantities 

 and the occurrence of the critical point. For purposes of com- 

 parison, of still greater importance than the melting points 

 would be the boiling points of the substances used as anode ; 

 but we have practically no data on this subject. 



The only metals showing the critical point well in free air 

 were iron and copper. Of these, copper has the lower melting 

 point and the better heat conductivity. Heat conductivity of 

 the metal, however, does not play the controlling part here, for 

 in the case of iron it makes but little difference in the occurrence 

 of the critical point whether the globule is formed on an iron 

 rod or placed in a poorly-conducting carbon holder. This is 

 due perhaps partly to low heat-conductivity of the oxide globule 

 itself, partly to its somewhat loose connection with the metal 

 behind it. 



We have found by tests with a blowpipe that copper oxide 

 melts at a lower temperature than magnetic oxide of iron. In 

 accordance with this, the critical point for copper comes at 

 smaller current and less expenditure of energy than that for iron. 



*Cf. Polak, Elektrot. Zeitsckr., xxiv, 599, 1907 ; Gekrke and Reickenheim, 

 Verh. d. Dentscken Pkys. Ges., v, 76, 1907.' 

 fPkys. Rev., xii, 137, 1901. 

 jwied. Ann., lxviii, 752, 1899. 



