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



coated with cement and ground plane. As the total width 

 o£ one strip and insulation was 1*57 mm., the lesistance per 

 metre of the ribbon was 0*6493 ohm, and the total resistance 

 of the, heater at the same temperature was 3*68 ohms., we 

 readily find that the energy developed per square centimetre 

 was 1*122 per cent, of the total energy. As the temperature 

 coefficient of nichrome is very small, errors due to the small 

 inequalities of temperature of the ribbon towards the edges 

 of the disk would be unimportant. 



In order to avoid trouble with leakage into the couples 

 alternating current was used. The energy was measured on 

 a wattmeter of the Siemens dynamometer type, which as 

 before controlled the relay lor maintaining a constant energy 

 supply. This instrument was carefully standardized on con- 

 tinuous current before use. 



The initial determination at about 121° 0. gave 4'67 x 10~ 3 

 and 3*54 X 10 -3 as the conductivities of granite and basalt 

 respectively. These results are some 12 to 15 per cent, lower 

 than the corresponding results by the other method, but very 

 approximately in the same ratio. There can be little doubt 

 that the low values are due to the films of cement across 

 which the heat-flow takes place. In the previous method the 

 flow was parallel to the cement film in all cases. 



Later results at higher temperatures apparently indicated 

 a fall of conductivity of both rocks, that for the granite being 

 much the larger. The results were rather inconsistent and 

 the last results at 115° C. were actually much lower than 

 those at 460° 0. which immediately preceded them. This 

 was almost certainly due to the failure of the cement, as on 

 examination it was found to have given way between the 

 rock disks. This failure must have been caused by differential 

 expansion, as in other parts the cement was apparently 

 unaffected by the heat. As cracking had probably started 

 in the cement some time before the final rupture, the results 

 are of little value, but indicate as far as they go that the 

 conductivity of granite is about 35 per cent, higher than 

 that of basalt at 124° 0. before heating to higher tempera- 

 tures, and falls much more rapidly as the temperature is 

 raised. 



As there seemed to be little chance of rinding any cement 

 which would withstand the stresses set up by differential 

 expansion, the rock disks were ground together with carbo- 

 rundum powder until they fitted extremely closely and very 

 fine grooves were cut in the outer faces of the inner rock 

 di>ks to accommodate the couples measuring the temperatures 

 of the faces between the rock disks, so that the rocks could 



