76 Mr. H. H. Poole on Thermal Conductivity and Specific 



passed. The P.D. wires lead in through grooves as shown 

 at A and B. The multiple couple was sunk in a series of 

 straight grooves as shown in the figure, in which half of the 

 couple is seen, the other eight junctions being cemented into 

 the opposite face. The grooves in the two faces were 

 arranged so that the inner junctions on one face were 

 opposite to the outer junctions on the other face and 

 vice versa. 



The transverse orooves in the two faces did not register 

 with each other but were about 2 mm. apart, thus decreasing 

 the chance of short circuit between different parts of the 

 couple. All these grooves were considerably finer than 

 shown in the figure, being about 0*6 mm. wide at the top. 

 The longitudinal couple grooves were about 1'2 mm. deep, 

 being V-shaped as before. The transverse and axial grooves 

 were about 0*3 mm. deep. The couple passed under the 

 central wire in one place as shown : here the transverse 

 groove was made deeper and the couple insulated from the 

 central wire by means of cement. The single couple is 

 shown at G. The couple leads along the outside of the 

 cylinders were only laid in their grooves, no cement being- 

 used. They were insulated from the copper furnace tube by 

 thin mica sheet which was cemented over the whole cylinder. 

 By this means all danger of these parts of the couples 

 breaking was obviated, and for the first time no trouble of 

 this kind occurred. The rest of the arrangements remained 

 as before. 



There are several advantages in dividing the rock longi- 

 tudinally, chief of which is the smallness and accurate 

 centering of the central aperture ; all danger of want of 

 uniformity of temperature of the central wire is eliminated 

 and the mean heat flux from a length of about G cm. of it 

 is taken. Another advantage is that the distance from the 

 junctions to the radial grooves is about 7*5 mm. instead of 

 1'4 mm., as was the case for the inner junctions in the 

 previous experiments. The use of eight pairs of junctions 

 doubles the E.M.F. to be measured for a given value of A, 

 which is also an advantage. 



The distances of the bottoms of the inner and outer 

 grooves from the centre being 0'493 cm. and 1*464 cm. 

 respectively and the distance between the P.D. wires 

 105 cm., we find that with the same notation as before 



K= 4-007 xlO~ 3 X -£-j 



where 



W r = W- 1-286 «. 



