Prof. Guthrie on the Conduction of Heat by Liquids. 285 



suredly are true of the other. The same identity may be predi- 

 cated with even greater safety of vapours and permanent gases. 



Magnus has endeavoured to show that hydrogen has a specific 

 and preeminent conductmg-power. 



The examination of the conducting-powers of liquids is mainly 

 due to Rumford, Murray, and Despretz. 



The well-known experiment of Rumford, in which water is 

 boiled over ice, led that experimentalist to conclude that water 

 does not conduct heat. 



Murray placed the bulb of a thermometer under a layer of oil 

 or mercury in a hollow cylinder of ice, and applied heat from 

 above. He concluded that both liquids conducted heat, and 

 that mercury did so more readily than water. 



Despretz* enclosed water in a wooden cylinder painted inter- 

 nally and placed thermometers down the axis of the cylinder. 

 He heated the upper surface of the water-column by means of a 

 copper vessel into which fresh portions of hot water were conti- 

 nually poured. After several hours (thirty or sixty) he obtained 

 a state of thermal equilibrium. Despretz concluded that in 

 water the temperatures of successive points of the axis of the 

 cylinder equally distant from one another are in geometrical 

 progression, and that accordingly the same law obtains with 

 liquids as with solids. The same observer also concluded that 

 salt dissolved in the water does not sensibly affect its conduct- 

 ing-power. Though Despretz inserted thermometers in the 

 wooden walls of his cylinder, and placed some also excentrically 

 in the water-column, and deduced that the column diminished 

 in temperature from the axis outwards and that the walls were 

 cooler than the column, yet, in such experiments, it is difficult to 

 admit that the conducting-power of the containing vessel, be it 

 greater or less than that of the water, is without influence. 

 Further, while the dimensions chosen by Despretz for the liquid 

 column (1 metre high, 405 millims. in diameter) are inadmissible 

 for many liquids, the thermometer-bulbs would interfere with 

 the thermal manifestations in smaller columns. 



In order to compare the conducting-powers of different liquids 

 with one another, and to measure approximately the absolute 

 conducting-power of one liquid, two conditions appear essential. 

 The liquid must be heated from above ; it must not be heated 

 in a vessel whose sides conduct with anything like the power 

 which the liquid enjoys. If the first of these conditions be neg- 

 lected, phenomena of convection directly intervene ; if the second, 

 such vitiating phenomena are brought about by the conduction 

 by the solid body. ; 



* Ann. de Chim. eide Phys. vol. lxxi. p. 206 (1839). Compies Rcndus, 

 vol. xxxiv. p. 540 (1852). 



