CONDUCTIVITY AND LUMINOSITY OF I'l.AMKs >\T.MMN<: V \h ii;]>r.l> BALTB 



ing two immiscible liquids of nearly equal specific gravity. A pressure of 1 centiin. of 

 water corresixmded to a motion of 10 centims. in tin- surface of separation. The 

 pressure actually employed produced u nnti>n of the surface of separation of 

 18'G5 centims. A constriction \\as ni.ul.- in tin- gas supply tul>e just In-fore it joined 

 tin- air, so that the pressure should !* great enough to IK; mc;isuivd with accuracy. 

 The variation in the gas pressure during the ex]>erimeiits did not exceed "25 per cent. 



It was found advisable to pass the mixture of gas, air, and spray through a 

 considerable length of apparatus, in order to allow only very fine sprav t-> get to the 

 flame, and to allow time for the thorough mixing of gas and air, on which the steadi- 

 ness of the flame so greatly depends. The arrangement of this part of the apparatus 

 is obvious from fig. 1. 



The cone-separating apparatus and electrodes are shown in tig. :\. The gaseous 

 mixture and spray pass up through the central tube, I, marie of thin brass. On this 

 tiilie the upper part of the apparatus could l>e slid up or down, so as to bring the 

 electrodes to any desired height above the mouth of the tube. The hard wood block, 

 B, was provided with three screws for centring the tul. A wider brass tul>o, O, was 

 fixed into the block, and was provided at the lower end with a cork, through which 

 the inner tul>e could slide. 



A large cork, 0, was cemented on to the wooden block in order to keep the glass 

 cylinder, G, in jxwition. The lower edge of the cylinder fitted into a groove in the 

 block and was trapped with mercury. 



A mica plate, M, kept in position by brass clips and pierced in the centre by a hole 

 2'2 centims. in diameter, was placed on the top of the glass cylinder. The electrodes 

 were supported l>y two rigid porcelain tilings, P, fixed into the wood block by a 

 packing of fusible metal. The electrodes consisted of two concentric cylinders of 

 platinum iridium alloy. The outside cylinder, E, was supported by a wire of the 

 same material, '5 millim. in diameter, thickened after a distance of T5 centims. to 

 2 millims.. and l>ent at right angles so that it reached for 3 centims. down the 

 porcelain tul)e, where it was joined to a platinum wire. By means of a cross-piece 

 fitting into a V-shaped groove at the top of the porcelain tube, this attachment of the 

 elertrode was kept in a fixed position. The inner electrode cylinder, I), was supported 

 in a similar way. The lower end of this cylinder was provided with conical cap so 

 as to avoid the formation of eddies in the gas stream. 



The dimensions of the electrodes were 



Height of cylinders 1 '575 centims. 



Inside diameter of outer cylinder '875 centiin. 



Outside diameter of inner cylinder .... '450 



The electrodes were set up so as to be concentric and co-axial with the flame, so 



that a symmetrical region it' flame gases was included. 



In our preliminary experiments \s.- tried eleetn-des consisting of platinum foil 



