268 THE ROYAL SOCIETY OF CANADA 



With these heavy currents special precautions had to be taken 

 to shield the arc against air currents and magnetic disturbances. 

 "Humming arcs," in which the luminous portion enlarges and con- 

 tracts periodically, and the voltage oscillates correspondingly, were 

 avoided by rasping off the tip of the cathode. "Groaning arcs," a 

 new form of unsteady arc, in which the white hot anode spot jumps 

 up and down between the bottom and the edge of the crater, were 

 avoided by cutting the electrodes to the proper shape, using a tem- 

 plate. 



The effects on current and voltage caused by raising or lowering 

 the cathode a few millimetres were measured; and from these results 

 by integration, the effect of the length of the arc. The "constants of 

 integration" were obtained by measuring clay models of the space 

 between the electrodes. 



In connection with these measurements it has been found, 

 contrary to expectation, that it is possible to maintain an arc with 

 a potential difference as low as twenty volts; in fact, the use of the 

 twenty volt arc has been introduced as part of the routine of building 

 up the current. 



The Melting Interval of certain Undercooled Liquids 



By Professor John Bright Ferguson 



Presented by Professor W. Lash Miller, F.R.S.C. 



In 1916, Mcintosh and Edson (J. Am. Chem. Soc. 38, 613), 

 reported that solids produced by the sudden chilling of certain aqueous 

 solutions did not behave on melting in a normal fashion. They 

 found that the melting points of this material corresponded to the 

 points on the liquidus curves on the phase rule diagrams of these 

 solutions. Constant temperature baths were prepared by them by 

 the use of such substances. 



We repeated some of their experiments in the hope of obtaining 

 some inkling as to the explanation of the phenomenon and also 

 because we did not believe that this method of obtaining a constant 

 temperature bath was of general application. We found that the 

 constant temperature obtainable with a given salt solution was 

 somewhat dependent upon the experimental proceedure and did 

 not always correspond to the liquidus temperature as they had held 



