136 Prof. J. Joly on a Method of Investigating tlie 



holder. Heating the retort for 40 or 50 minutes sufficed to 

 fully complete the decomposition of the rock. There was 

 difficulty experienced in obtaining a quite air-tight joint at 

 the junction of the head and body of the retort. 



Several experiments were subsequently made on the possi- 

 bility of breaking up the powdered rock in the electric arc 

 without the intervention of chemicals. But difficulties arose 

 in securing complete volatilization. The action of the arc 

 largely resulted in scattering the powder. For this and 

 other reasons the method was abandoned as less reliable and 

 controllable than the collection of the emanation during the 

 decomposition of the rock in presence of chemicals. This 

 latter method I determined to develop in connexion with the 

 advantages arising from the use of the electric resistance 

 furnace. 



Two forms of this furnace have been tested. One of these 

 is available for use with an ordinary platinum crucible ; the 

 other — of the tube-furnace pattern — requires the use of a 

 special platinum boat. I shall first describe the former. 



A disk-shaped base of iron has an annular groove turned 

 in its upper surface. This groove receives a silica beaker or 

 shade which is inverted over the crucible. Mercury poured 

 into the groove renders the interior of the silica shade air- 

 tight against small changes of pressure. The volume of air 

 included beneath the shade amounts to about 500 or 600 c.c. 

 The accompanying figure is to half-scale, so that the dimen- 

 sions are readily obtained from it. The furnace is shown in 

 vertical section. The disk is supported on short iron feet, 

 and during an experiment stands in a vessel of water which 

 maintains its temperature below that of the ebullition of 

 water. 



The platinum crucible is contained in an outer crucible of 

 Berlin porcelain, a layer of magnesia powder intervening. 

 In the magnesia powder the platinum resistance wire, which 

 is heated by the current, is inserted. It is folded in a par- 

 ticular manner so as to supply the heat where most serviceable. 

 A ring of asbestos card covers in the magnesia and supports 

 a platinum cover or shield to reflect back heat and stop the 

 spluttering of the melt injuring the clear silica shade. The 

 outer porcelain crucible rests on several thicknesses of asbestos 

 card contained in a recess on the surface of the iron disk. 

 Beneath the supporting asbestos a perforation of the disk 

 permits the escape of the evolved gases through a tube 

 leading from the under side of the iron base. A groove is 

 cut in the asbestos cards in such a manner as to leave a clear 

 opening leading to the escape pipe ; without this there is 



