Conduction of Heat in Liquids. 3 



instituted in the upper cone, and the time -interval observed 

 when an alteration in the level of the coloured fluid became 

 apparent. This he denotes by t, and calls the " time-of-heat 

 penetration." He denotes by T the temperature of the lower 

 cone, by dT the excess of temperature of the current in the 

 upper cone, and by 8 the thickness of the liquid layer. He found 

 that t diminishes when dT is increased, and is not directly pro- 

 portional to 8. He thence unwarrantably assumed the con- 

 ductivity to increase with the temperature. Possibly a clear 

 idea of his views will best be given by a statement he makes 

 on page 649 : — " If we had found that the time £ . . . had been 

 the same for heat of all temperatures, we should of course 

 expect to find the time t proportional to the thickness 8." A 

 very slight acquaintance with the mathematical theory is suf- 

 ficient to show that this is incorrect. In fact there should be 

 an immediate effect gradually increasing ; and the interval, 

 prior to its detection, depends on the delicacy of the apparatus 

 and the fineness of the observer's senses. 



In a second series of experiments he obtained what he terms 

 a " measure of the resistance " of the conductivity. First, 

 the bases of the cones were placed in contact, and the level of 

 the coloured fluid read off at equal intervals of time subsequent 

 to the commencement of a hot current in the upper cone. 

 The cones were then separated to a certain distance, a liquid 

 introduced, and the experiment repeated. The difference 

 between the readings in the two cases he considered a direct 

 measure of the quantity of heat stopped by the liquid, and this 

 he supposed to measure the thermal resistance, i. e. the reci- 

 procal of the conductivity. On page 658 he says : — " . . . the 

 specific heat of the liquid in these contact experiments is of 

 no influence;" and on the following page he shows how to 

 deduce numerical results from the observations. 



The theory is altogether faulty, and the numerical results, 

 though doubtless accurate measures of something, are totally 

 misleading. I have been unable to find any means of obtain- 

 ing probable results from the data supplied. 



There are two points, however, of great importance that 

 Prof. Guthrie's experiments seem to establish. Apart from 

 conduction proper or convection-currents, there are still two 

 ways in which heat might travel from the upper cone to the 

 lower. This might take place by radiation ; and that a certain 

 amount of heat is so transferred is unquestionable. That it 

 is comparatively small may be deduced from the fact that a 

 decided interval (several seconds) elapsed after the heat was 

 applied before any movement of the coloured fluid could be 

 detected. Prof. Guthrie was evident] v much influenced by 



B 2 



