Titanium, and Vanadium by Thermelectronic Currents. 247 



that not more than 4 seconds are needed for securing a well- 

 developed spectrographic record at 3000° C. 



All experiments were made in air at atmospheric pressure, 

 and a fresh plate of graphite was used for every observation. 



§ 3. Temperatures of graphite plates for various potential 

 gradients of heating current. 



Temperature determinations were made by means of a 

 Wanner pyrometer. In one series of readings — Nos. 1-12, 

 the instrument was directed upon the underside of the 

 heated plate at an angle of about 45°, the slit being about 

 7 cm. from the plate. In another series — Nos. 13-34, the 

 graphite plate was fixed in the space between the coreless 

 field coils of a Faraday electromagnet and the hot surface 

 was sighted through one of the coils. The distance between 

 plate and pyrometer was in this case about 16 cm. A new 

 plate of graphite was of course used for each reading and 

 the effective length of the plates was kept as nearly as 

 possible at 27*5 mm. throughout. A voltmeter reading was 

 taken in each case, but an ammeter reading only for 

 Nos. 13-34. 



On account of the short life of the plates, especially at 

 the higher temperatures, and also in view of the fact that 

 the resistance of the plates increases very rapidly after a 

 heavy current has been passing through them for only a few 

 seconds, all the necessary manipulations, settings, and readings, 

 involved in the operation of determining the plate tempera- 

 ture, had to be accomplished within 10 to 15 seconds. The 

 procedure adopted was generally as follows : by means of 

 the adjustable resistance the current was at first raised 

 slowly until the plate had become hot enough to be observed 

 through the pyrometer, previously set for a low temperature. 

 The current was now brought rapidly to the desired value, 

 whilst at the same time the pyrometer was manipulated and 

 adjusted in keeping with the rise in temperature. As soon 

 as a steady state was reached (as indicated by the voltmeter) 

 and the pyrometer finally set, a voltmeter reading was taken. 

 The pyrometer was read only after the end of the experi- 

 ment. When the determination of the plate temperature 

 was made in a magnetic field the latter was put on first, and 

 remained on throughout the experiment. The strength of 

 the magnetic field used was 5y0 c.G.s. units and readings 

 w r ere taken with the magnetic force acting both upwards 

 and downwards upon the current passing through the plate. 



The following table contains the results of all the observa- 

 tions made, even those which are obviously out. On the 



