510 



Professor W. ChamUer Roberts- Austen [March 15, 



It is only necessary for our purpose to use a portion of the long 

 scale, which may be traced across the end of the room by the spot of 

 light from the galvanometer, but we must make that portion of the 

 scale movable. Let me try to trace before you the curve of the 

 freezing of pure gold. It will be necessary to mark the position 

 occupied by the movable spot of light at regular intervals of time 

 during which the gold is near 1045° C, that is, while the metal is 

 becoming solid. Every time a metronome beats a second, the white 



l^^J 



Fig. 



screen A (Fig. 5), a sheet of paper, will be raised a definite number 

 of inches by the gearing and handle B, and the position successively 

 occupied by the spot of light C will be marked by hand. 



You see that the time-temperature curve, x y, so traced is not 

 continuous. The freezing point of the metal is very clearly marked 

 by the vertical portion. If the gold is very pure the angles are sharp, 

 if it is impure they are rounded. If the metal had fallen below its 

 freezing point without actually becoming solid, that is, if superfusion, 



