ON THE THERMAL CONDUCTIVITIES OF CERTAIN ROCKS. 127 



conductivities which, on account of their relative values only, could not 

 receive admission to the store of absolute data furnished by Professor 

 Everett's descriptions and reductions. In relation to this procedure the 

 example of Peclet has been followed in the present list, who, in 1841, 

 adopting relative numbers found by Despretz in 1821 and 1827, assigned 

 absolute values to Despretz's series, by means of his own absolute deter- 

 mination of the thermal conductivity of one of the substances (statuary 

 marble) examined by Despretz. The values so assigned conjointly by 

 Peclet and Despretz are marked ' Despretz (a) ' in the Table. But in a 

 later relative list of Despretz, of a few specimens of rocks tested in 1852 

 (not used, apparently, by Peclet in his general list of 1853), the value 

 given for marble, for reduction of Despretz's earlier list, is not adhered 

 to in the present Table, but the average conductivity found by a number 

 of experiments of the Committee on different marbles is used instead of 

 it, and Despretz's relative values thus converted into absolute ones are 

 marked ' Despretz (/3), 1852,' in the present list. 



A similar course had to be pursued in the case of two important 

 relative researches published almost simultaneously in 1856 and 1857 

 by Helmersen of St. Petersburg, and Hopkins of Cambridge, on 

 thermal rock-conductivities. To the relative quantities given by those 

 researches for vein-quartz and white statuary marble respectively, the 

 average values of the absolute conductivities of those rocks found by the 

 Committee's repeated observations of them, are assigned (with a slight 

 deviation in the latter case for the reasons recognised below), to convert 

 the other relative quantities of the two researches into absolute measure, 

 and to allow of their being compared on a common scale with other absolute 

 determinations in the general list. 



The work of Helmersen terminated in recording the observed tem- 

 peratures along bars of the tested specimens one and a-half inch 

 square and eighteen inches long, heated at one end, according to 

 Despretz's method. Pour thermometers being distributed at equal 

 distances along the bars, it was found that in the worst-conducting bars 

 the logarithmic decrement of temperature was so far from uniform and 

 constant, that abandoning the doubtful measurement of its true value in 

 each of them, Helmersen was contented to conclude from his experiments 

 that vein-quartz was the best conducting material of those which he 

 examined. 



As, however, unusual care was bestowed on Helmersen's experiments, 

 the relative quantities which they denote are yet of great value for 

 corroborative estimations. The decrements for each bar, and for each 

 interval between the four thermometers, was therefore deduced from 

 them, and by a mode of reduction to a mean decrement for each bar used 

 independently of each other by two of the Committee to reduce the 

 observations with the best presumptive regard to their respective weights, 

 to relative conductivities, the following two series of numbers were 

 obtained : — 



Mica Serpen- Sand- Lime- 

 Rock Specimen Quartz Schist Granite Marble Porphyry tine stone stone 

 Kelative ] I. 00710 -00501 00356 00510 -00448 -00555 -00471 -00466 

 Conductivity, I 



7e jll. -00688 -0056.^ -00375 -00412 -00371 00458 00508 00399 



Of these two scries preference has been given to the latter, as it gives 

 more weight to the thermometric intervals near the source of heat, that 



