138 Mr. W. L. Upson : Observations 



of the negative. The arc in which carbon forms both posi- 

 tive and negative partakes of the properties of both of these 

 groups, occupying its individual position under the influence 

 of the carbon negative ; but being little influenced by its 

 positive, it keeps the same general shape as the lower group 

 of curves. 



Fig. 6 gives this same group of curves for arcs in 

 hydrogen. The metal-carbon and carbon-carbon arcs so 

 nearly coincide that one curve is given to represent them. 

 The real positions of this group may be seen in fig. 10. We 

 now have not a difference in curvature, for the curves are 

 approximately parallel, but merely variations in the distance 

 apart, according to the material of the negative. The posi- 

 tion of the curve, therefore, seems to be due to some constant 

 of the material — say, its ability to become ionized in the 

 particular gas. 



In air there is superposed the active effect of oxidation. 

 We may infer that the general shapes of the curves, — or, at 

 any rate, variations in shape — are due to the influence of the 

 surrounding gas. 



Fig. 7 shows the characteristics of arcs in air when 

 both electrodes are of the same metal. That of aluminium 

 differs considerably from the other two, and in general 

 it was found that where aluminium entered, either as 

 one or both of the electrodes, there was apt to be consider- 

 able variation, whereas iron and copper kept pretty close 

 together. 



Fig. 8 gives the combinations of the different metals 

 in air. Fig. 9 gives them for arcs in hydrogen. It has 

 already been mentioned that these arcs bear more resem- 

 blance to heavy spark discharges than to genuine arcs. But 

 the action seems to follow the same general law, in both 

 cases. Home interesting comparisons may be made from 

 these curves and those of figs. 5 and 6. 



In figs. 8 and 9 we find the highest and steepest curves 

 to be Cu-Al, and C- Al occupies a similar position in 

 fig. 5. In figs. 8 and 9, again, the lowest and flattest 

 curves are for Al-Fe. The copper and iron curves bear a 

 general resemblance to each other throughout, but with a 

 certain peculiarity. 



We have considered the relative height of the curves to be 

 due, in part, to the effect of the surrounding gas, and in part 

 to a specific property of the material of the negative elec- 

 trode. There may be other influences, but these two are 

 suggested by a comparison of the copper and iron curves in 

 figs. 5, 6, 8, and 9. The curves both for Cu-Fe and C-Fe 



