Fletchioij — The Melting-Poinfs of some of the Rarer Minerals. 445 



along the axis of a tubular draught-shield E. A slight tension is exerted on the 

 forceps by the spring at D, sufBcient to keep the ribbon taut without producing 

 permanent extension at unduly low temperatures. One forceps is produced 

 in a flexible arm (?, which by completing an auxiliary circuit at 

 tlie point P of a graduated micrometer screw H serves to record the 

 extension, and hence tlie temperature of the strip. A small electro-magnet Ji" 

 indicates by an arm, which enters the field of vision through a slot in the 

 tube of the microscope, the precise moment at which the auxiliary circuit is 

 complete and contact made. The great advantage of this nietliod of 

 temperature measurement is that, by following up the flexible extension 

 of the forceps with the micrometer screw, the temperature at the exact 

 moment required is recorded ; and sudden small changes in the temperature, 

 brought about by drauglits or small variations in the resistance, are very 

 easily remedied by a rapid motion of the screw-head. The needle of the 

 electro-magnet is dead beat, and therefore facilitates rapid and accurate 

 measurement. 



A. particle of the substance to be examined is put near the centre of the 

 platinum, and observed by the low-power microscope shown, wliich lias a 

 motion parallel to the strip. ^ 



Curve of Thermal Extension of Platinum Ribbon. 



With regard to the curve of thermal extension upon which the temperatures 

 are determined by interpolation (fig. 3), the points were decided only from the 

 results of the most consistent experiments. It is not necessary to plot a fresli 

 curve for each new platinum ribbon as it is usually found that they are 

 coincident from one ribbon to another, when care is taken to cut each fresh strip 

 after the exact pattern of the standard on which the curve was originally 

 determined. On using a new ribbon it was found sufficient for the purpose 

 of verification to re-determine two points by melting, say potassium carbonate 

 and metallic palladium. The curve reproduced shows the increase in the 

 coefficient of expansion witli temperature, and the evenness of its slope is some 

 additional assurance of its accuracy. It was not considered necessary in these 

 experiments to make allowances for diurnal differences in the temperature 

 of the laboratory. 



Tiie standard determinations should be made upon purified substances ; 

 ordinary laboratory chemicals will be found to give fusion points differing 

 from those of the purified samples. 



The following substances of guaranteed purity were used in determining 



3x2 



