Cady and Arnold — Electric Arc. 393 



energy at the beginning of the second stage, that the galvan- 

 ometer deflection decreased but little. In so far as the junction 

 can be considered as giving the mean temperature of the 

 globule, this means that the globule temperature is not propor- 

 tional alone to the energy expended, but tends to become con- 

 stant when the second stage is reached. 



§14. Heating the Positive Globule. — Anything that raises 

 the temperature of the positive globule should cause the criti- 

 cal point to be reached at a smaller current. This we have 

 found to be the case. With thin electrodes, which conduct 

 the heat away less readily, the critical point comes at a slightly 

 smaller current than with thick electrodes. Also, the smaller 

 the positive globule, the sooner is the critical point reached. 

 Thus with the same anode, we have recorded within a few 

 minutes values of the critical current from l'l to 1*7 amperes, 

 depending on the size and curvature of the positive globule. 



The experiment was also tried of heating the positive termi- 

 nal, an iron rod 3 mm in diameter, by means of a blast lamp. A 

 large mica shield fitted closely around the rod about a centi- 

 meter below the arc, to protect the latter from the direct effects 

 of the flame. When the positive terminal was heated to 

 redness, the critical current was 1*4 amp. Without artificial 

 heating, the critical current was 1*7 amp. The globule was of 

 approximately the same size in each case, and the blast lamp 

 was kept burning throughout the experiment in order that the 

 nature of the gas surrounding the arc might be constant. 



Stark and Cassuto (1. c.) have shown that heating the anode 

 causes the positive drop to increase slightly, owing probably to 

 increased thermal electromotive-force. This, by increasing the 

 expenditure of energy at the anode, would also tend to cause 

 the critical point to come at a smaller current. 



When it was attempted to cool the anode by surrounding it 

 with snow, anomalous results were obtained, which we found 

 to be due to the effect of the water vapor. For example, when 

 steam w T as allowed to flow freely into the inclosure surrounding 

 the arc, the break from the first to the second stage took place 

 at a much higher voltage and smaller current than when the air 

 was not thus saturated with moisture. The difference — amount- 

 ing to about ten volts in the case of the potential difference — is 

 too great to be accounted for by change in the heat conductivity 

 of the air due to the presence of moisture. It is more likely 

 that the water vapor, which as the experiments of Merritt and 

 Stewart* indicate becomes highly ionized in the arc, gives to 

 the arc quite different characteristic curves from those prevail- 

 ing in ordinary air. 



* Phys. Eev., vii, 147, 1898. 



