Machine for producing High Tension Electricity. 179 



it some very interesting results. A drawing of it accompanies this 

 paper (Plate 2). 



My attention was at first directed to the heating effect of the 

 secondary current at an air gap on its circuit. The amount of heat 

 developed at this point when all the coils were in operation proved 

 unexpectedly large, and I was surprised to find it almost entirely con- 

 fined to the negative side of the gap. I found, also, that when the 

 current passed in sparks very little heat was exhibited. It was only 

 when the passage of the current assumed the appearance and condi- 

 tion of an arc that the heat came into prominence. Taking platinum 

 wires of No. 27 B.W. gauge for the positive and negative terminals, 

 which I shall call electrodes, to distinguish them from the sparking 

 points at the collector, I found that the ends could be drawn asunder 

 to a distance of fully an inch before a decided stream of crackling 

 sparks was elicited, and. even then the sparks presented a hazy 

 appearance. On re-advancing the eiids to each other the sparks 

 diminished, while the haze increased and gradually assumed the con- 

 dition of a pale blue arc. At a distance of 0'6 inch a well-defined 

 arc could be maintained, though not without a slight admixture of 

 faint sparks, which followed the curvature of the arc. At this dis- 

 tance the heat was sufficient to fuse the end of the platinum wire 

 forming the negative electrode, but the heat did not reach its maximum 

 until the separating distance was reduced to a few hundredths of an 

 inch. At that small distance the negative platinum melted with 

 great rapidity, and ran back in a globule until it got out of melting 

 range of the arc. By following up the melted globule by steadily ad- 

 vancing the positive wire, the negative wire fused at the rate of 

 nearly 3 inches per minute ; but, however long this process was con- 

 tinued, the positive wire remained to all appearance perfectly cool. 

 Retaining the positive wire unaltered, I increased by successive steps 

 the thickness of the negative wire, to determine the point at which it 

 became too thick to melt and run back in a globule, and it was not 

 until I increased the thickness up to 16 gauge, which corresponds to 

 a diameter of O065 inch, that I reached this limit. With that wire 

 the end melted into a rounded form, but no longer receded in a 

 gathering globule, as thinner wires did. Iridioplatinum wire of 21 

 gauge, containing 70 per cent, of iridium, readily melted and receded, 

 though not rapidly, until the globule fell by its weight. In this 

 instance the light emitted from the globule was so intolerably bright 

 as to require darkened spectacles to view it with impunity. I next 

 proceeded to ascertain how far it was necessary to reduce the thick- 

 ness of the positive wire before it exhibited an equal degree of heat 

 with the thickest wire opposed to it on the negative side, and I found 

 that I had to diminish its diameter to 31 gauge before this equality 

 was reached. Now, the sectional area of No. 16 gauge is 42 times 



