320 Mr. John Allen Harlcer [Feb. 9, 



up to sun's temperature. If we know the law of variation we can 

 measure the temperature by the use of some kind of photometer — 

 which is what all optical pyrometers are. 



For obtainino; really high temperatures electric furnaces are our 

 only resort. Small gas furnaces can reach 1600° with difficulty; 

 large industrial furnaces attain 1800° C. in some instances. 



Mr. Cook, of Manchester, has kindly lent me for this occasion a 

 number of electric furnaces. These are constructed by winding tubes 

 of fire clay or alumina with nichrome or platinum wire : the external 

 lagging is of kieselguhr. Steady temperatures up to about 1000° 

 and 1200° C. respectively can readily be got with power from a com- 

 mercial circuit of 100 or 200 volts. Witli thicker wires and current 

 at lower voltage these upper limits can be appreciably extended. 



For higher temperatures we have to make use of carbon or 

 graphite, and electric heating wa'^ first applied by such means in the 

 form of the arc furnace. Such a furnace has many inconveniences— 

 the heating is intensely local, and there may, for example, be a 

 gradient of 2000° C. in a single inch. There is practically no tem- 

 perature control, and there is every possibility of the final product 

 becoming largely contaminated with carbon. Most of the early 

 isolated so-called elements have since proved to be largely carbides. 



Resistance heating is usually much more convenient, and this is 

 the principle of carbon-tube furnaces, some essential features of which 

 were employed by Prof. Dewar many years ago. They will stand 

 rough use, and are much more controllable than the arc furnace. It 

 is as easy to control a temperature of 2500° C. as one of a red heat. 



Such furnaces usually have their end-terminals water-cooled, and 

 are surrounded by lagging of lamp-black or charcoal. 



The furnace tubes are either straight if made of carbon (Fig. 2), 

 or spiral if made of Acheson graphite (Fig. ?>). In the latter case 

 they are provided with an internal liner-tube of carbon. There is no 

 special difficulty in cutting the spirals from the solid ; graphite, un- 

 like amorphous carbon, is an extremely tractable substance to machine. 



We have used these carl)on resistance furnaces a great deal at the 

 National Physical Laboratory, and Mr. Greenwood, at Manchester, 

 carried out his experiments on boiling metals by the aid of such a 

 furnace. The boiling of a metal (shown) forms a not-impossible 

 lecture experiment, and a projected image of the surface of boiling 

 tin displays all the usual phenomena of ebullition. The heating up 

 of carbon is somewhat strikingly shown by passing a heavy current 

 through a thin broad carbon strip provided with water-cooled ter- 

 minals (experiment shown). The stream lines of heat flowing from one 

 terminal to the other are well illustrated at one stage of the heating. 



Among other methods of electric heating are the induction furnace, 

 which is of great value in refining crude materials, and the flame 

 spark, in which it is possible to volatilize as refractory a substance as 

 an incandescent gas mantle. 



