484 



Mr. J. E. Petavel. An Experimental 



could be slid along the U so as to increase or diminish the effective 

 length of the mercury. The clips Ci and Co were hollow and kept 

 cool by a water circulation. A rapid flow of water was also maintained 

 over the surface of the mercury. The electric circuit was completed 

 through the mercury switch S and an amperemeter. 



Fig. 5. — Apparatus for the fusion of platinum bars by an electric current. 

 A, platinum bar ; Cj, C 2 , copper clips in which the platinum is held ; L, lime 

 trough which supports the fused metal; R, mercury trough serving as a 

 variable resistance ; D, copper short-circuiting piece ; S, mercury switch ; 

 Tj, T 2 , massive copper leads to the battery terminals ; I and O, water inlet and 

 outlet. 



With this arrangement currents up to 2000 amperes could be 

 maintained for several hours. The platinum was supported by a 

 trough, L, of some refractory material which was carefully cut to fit the 

 metal bar. Erom the first it was evident that the principal difficulty 

 would be to find a suitable material for this support. The weight of 

 the molten metal acting on the supporting material gradually causes 

 the shape of the trough to alter, and in a comparatively short time the 

 cross-section of the metal becomes very irregular. The platinum then 

 superheats in the place where the cross-sectional area has been most 

 reduced, a spark occurs, and the molten metal thrown aside by the 

 explosive force of the discharge, freezes before it has time to flow back 

 into position. It is of course easy to re-weld the platinum, but the 

 same difficulty will recur time after time and at more frequent inter- 

 vals as the shape of the trough becomes more and more distorted. 

 Some twenty bars of different shapes and sizes were used with the 

 same result. Had it been possible to use larger bars, any small 

 distortion of the trough would probably not have affected the 



o 



