ON THE THERMAL CONDUCTIVITIES OF CERTAIN ROCKS. 141 



difference of thermo-currents observed periodically while cooling. (Sum depen- 

 dent on h, difference on k.) For bad conductors, cubes or spheres with central 

 and external thermo-pairs similarly treated. (To convert units to Peclet's, multi- 

 ply by •0848 ; to Ingstrom's, by -0509.) 



Oil 1-37 Cu 1306 



Sulphur 1-68 Brass 356 



Ice 13-5 Zn 362 



Snow 4-2 German silver .... 129 



Frozen soil .... 108 Fe 193 



Sandstone 



Large-grained granite 

 Serpentine (Zoplitz) 



Angstrom, ib. 67, 379 (1863). — Ear, exposed at one section to periodic heating and 

 cooling, and K deduced from observations of periodic temperatures at other 

 portions. Cu 54-62, Fe 9-77 (grm. cm. min.). 



Jannettaz, ib. (4) 29, 5 (1873). — Crystals. Method essentially de Senarmont's, 

 but modified and improved in details (plates of apparatus given). Confirms, 

 on a large number of crystals, the general results obtained by de Senarmont as 

 to relation of axes of conductivity to crystallographic axes and cleavage. 



T. Thompson, 'Mch. Journ.' 1, 81 (1802). Eegarding convection. Experiments to 

 prove that the motions of amber in Eumford's experiment do not imply currents 

 in the fluid. Two layers of fluid, coloured and colourless, superposed and con- 

 taining solid particles. On heating, the solid particles travel beyond the coloured 

 into the colourless fluid, without being accompanied by coloured fluid. 



Murray, ib. 1, 165, and 241. Conduction in fluids. Eumford's experiments with 

 heating from below inconclusive, prove only that not all the transference of heat 

 in a fluid is conduction. Heating from above, by hot oil poured on water, or 

 brass ball immersed in it over thermometer-bulb, tliermometer rose. But this, 

 perhaps, partly from conduction by sides. That sides do conduct shown by 

 jacketing cylinder with outer one filled with water and containing thermo- 

 meter, outer thermometer likewise rose. To obviate errors from this, made 

 vessel of ice with thermometer frozen through side, filled with oil or Hg, and 

 warmed by hot water poured into can floating or suspended on surface of liquid. 

 Thermometer rose in all cases. Not radiation, because when hot water sus- 

 pended just over surface of fluid without touching, rise of thermometer exceed- 

 ingly slight. 



Traill, ib. 12, 132 (1805). — Fluids. Turned wood vessel, with thermometer, and lid 

 bored with hole to admit hot iron cylinder, resting by flange on edge of hole. 

 With different liquids, thermometer took different times to rise 3° ; had sides of 

 vessel alone operated, times should have been equal. 



Rumford, ib. 14, 353 (1806). Heats water-surface from the point of a cone just 

 dipping into it, and fails to raise the temperature of a thermometer beneath. 

 Admits, however, that thermometer might have been raised, had heated particles 

 of water been prevented from rising up outside of the cone. 



Langberg, ' Ph. Mag.' (3), 20, 161 (1846).— Metals, Cu, Steel, Sn and Pb. No ther- 

 mometers, but thermo-pair applied to surface. Bars (or wires) about 36 mm. 

 long, and 1-7 x 1 mm. section. Law of Biot found not to hold. But ? as to 

 correctness of thermopile indications, from thinness of bars and long contact. 



Wiedemann, ib. (4) 10, 393 (1855).— See ' A.C.P.' (3), 45, 377. Longer and more 

 detailed abstract here. 



Thomson, ib. (4) 22, 23 and 121 (1861).— Eeduction of underground temperature 



observations, and deduction of K for the rocks in the three Edinburgh stations. 



See Trans. R.S.E, xxiii. (1860) p. 426, and these Reports (on 'Conductivity ') for 



1875. 

 Neumann, ib. 25, 63 (1863).— Translation from * A.C.P.' (3), 66, 183. 

 Angstrom, ib. 25, 130. — Cf. Pogg. 114, 513, of which this is a translation, fuller than 



the abstract in ' A.C.P.'(3) 67, 369, In addition to figures for Cu and Fe, deduces 



