FURNACE REACTIONS UNDER HIGH GASEOUS PRESSURES. 445 



material splits when subjected to shock there is little doubt that each minute cavity 

 still retains gas under a considerable pressure.* 



The small weight of vitrified product is ascribed to the high cooling effect of 

 hydrogen, but the properties of the material were similar to those noted in the earlier 

 work. 



Finally, an attempt was made to attain the highest {>o88ible temperature by 

 concentrating the heat of a powerful arc in the centre of a mass of sand, the 

 equipment of the furnace being of the type shown in fig. 11, D. 



No difficulty was experienced in maintaining an arc of some 10 kilowatts for an 

 hour or more, the electromotive force rising to 120 volts under a pressure of 180 

 atmospheres. 



A hollow sphere of 18 centims. external diameter and weighing about 5 kilogs. 

 was obtained, which, however, was no more transparent than with the less intense 

 heating. 



ON THE FORMATION OF CARBORUNDUM. 



In 1893 ACHESON discovered that a mixture of carbon and silica heated around a 

 resistance core readily produces a carbide of silicon to which the name of carbo- 

 rundum was given. Its preparation is now carried out on a large scale. It was 

 therefore of some interest to study the reaction, first under the usual conditions, then 

 under high pressure. 



We have prepared some quantity of this material in the laboratory and found it to 

 differ from the commercial product only in the smaller size of the crystals. The 

 amorphous variety invariably occurs in considerable amount surrounding the crystal- 

 line layers.t 



In carrying out the reaction in the pressure furnace the mixture of sand and carbon 

 was disposed around a core of granular retort carbon as shown in fig. 11, B. 



To insure a uniform cross-section this resistance core was tightly packed into a 

 thin brass tube which was placed in the axis of the furnace and supported between 

 the two graphite electrodes. As the mixture when heated becomes somewhat 

 conducting, it is advisable to line the furnace with a thin insulating layer of sand or 

 other material in order to avoid a short circuit through the iron. 



In these experiments the gas generated by the reaction was retained in the furnace 

 and the pressure allowed to accumulate up to some fixed limit (100 atmospheres). 



A typical example of such a run is given in Table VI. As will be seen, the brass 



* Relative to this subject an interesting investigation has been carried out by ARTHUR L. DAY in 

 America. Using lower pressures, which, however, were only applied after a high temperature had been 

 attained, he has prepared a glass containing relatively few air bubbles. ' Science,' N.S., 1906, vol. 23, 

 pp. 670-672. 



t See also S. A. TUCKER and A. LAMPEX. ' Journ. Amer. Chem. So<.-.,' 1906, vol. 28, pp. 863-868. 



