Eesearch on some Standards of Light. 



485 



experiments. However, as already stated, 2000 amperes was the 

 maximum current available, and with this current it was found 

 impossible to melt bars of much more than 70 sq. mm. in cross- 

 section. 



The length of the span between the edges of the clips was usually 

 70 mm., but in some cases this length was reduced to 40 mm. 



Lime, alumina and magnesia were the materials used for the 

 troughs. 



It became evident that small platinum bars could not be maintained 

 for any length of time in the molten state when supported on the 

 ordinary refractory materials, and it was decided to make use of metal 

 supports. 



The question of cost excluded the use of the metals of the platinum 

 group, and in default of a better material the trough was made of 

 nickel. A number of nickel plates insulated from each other by mica 

 were bolted together so as to form a solid block. In this block a 

 cylindrical groove was cut, its axis being perpendicular to the direction 

 of the laminations. The nickel plates were 1*8 mm. in thickness, 

 every alternate plate projecting 5 mm. on the sides and 6 mm. below. 

 Through the channels thus formed a rapid circulation of water was 

 maintained. The platinum bar was semi-circular in cross-section and 

 deeply grooved on the lower side where it came in contact with the 

 nickel. The shape was much the same as would be obtained by cut- 

 ting an ordinary bolt in two along its axis, the edges of the V-shaped 

 thread bearing on the nickel. In this manner the somewhat anomal- 

 ous experiment of fusing platinum in a nickel crucible was successfully 

 carried out. Needless to say that the anomaly is only apparent, the 

 molten platinum did not come in contact with the nickel, but was 

 supported in a shell of solid platinum. The cooling was so efficacious 

 that the nickel crucible never once became red hot. When the metal 

 is fused in this manner the quantity of heat lost is very considerable, 

 and the effective cross-section of the bars had to be greatly reduced. 

 The cross-sectional area, as measured from the bottom of the grooves, 

 did not exceed 30 sq. mm. At the surface of the bar the central 

 channel of molten metal was not more than 2 mm. in width. If the 

 current were forced so as to melt a larger surface, the bottom of the 

 threads softened sufficiently to allow the fused platinum to run 

 through, thus causing the bar to break. 



Even when this occurred the platinum was found to freeze rapidly 

 enough to prevent it alloying with the nickel support. 



If the above experiments were made on a larger scale some valuable 

 results would certainly be obtained. By the use of a large welding 

 transformer there should be no difficulty in fusing bars of ten or 

 twenty times the cross-section of those used in the present instance. 

 As will be seen below, good results can be obtained when the ordinary 



