616 



BRIDGMAN. 



that this region of imperfect contact was an undesirable feature^ 

 because it might give rise to sHght changes of resistance under pressure. 

 Very good connections were finally made by using pure gold as a 

 solder. A wire of platinum and molybdenum were laid side by side 

 and wrapped with pure gold wire, 0.004 inch diameter. By arcing 

 in hydrogen from a graphite point to the end of the wire the gold 

 melts and runs back over the surface, wetting both platinum and 

 molybdenum, and making a very good contact. 



Before the final runs on molybdenum, a complete set of runs was 

 made with the molybdenum wound on a core of hard rubber instead 

 of bone. There were large initial irregularities due to the expansion 

 of the rubber, which disappeared above 1000 kg. Between 1000 and 

 12000 kg. the results of the first run agreed within the limits of error 

 with those of the final runs. 



The smoothed results are collected in Table XVI and the experi- 



TABLE XVI. 



Molybdenum. 



mental values of mean coefficient and deviation from linearity in 

 Figure 18. With the exception of three bad points at 100°, the larg- 

 est departure of any single point from the smooth curve was 0.4% 

 of the total pressure effect, and the average numerical departure was 

 0.075%. The deviations from linearity are within the limits of error 

 symmetrical and parabolic; there is no need for graphical representa- 

 tion. 



