40 PROFESSOR A. CRICHTON MITCHELL ON THE 



follows a different law which Newton has not taken into consideration. That Fourier 

 correctly appreciated the importance of the condition attached to Newton's Law of 

 Cooling, viz., that the body was placed " non in sere tranquillo, sed in vento uniformiter 

 spirante," is seen from the exceedingly careful manner in which he defined # the 

 coefficient h, the ' conductibilite exterieure ' of a body. 



But the long succession of commentators, from Martine onwards, have practically 

 criticised Newton's Law on the assumption that the cooling body was placed in 

 (so-called) still air. Dulong and Petit only refer to its most important condition as 

 the action of a constant cooling cause. 



So far as I am aware, Leslie t was the only one who attempted to realise experi- 

 mentally the conditions required in any investigation whose object is to determine the 

 accuracy or otherwise of Newton's Law of Cooling. A metallic vessel, containing water 

 at a temperature above that of the air, was tied to the end of a string and whirled in a 

 circle for a definite period of time, after which the diminution in temperature of the 

 water was noted. From an experimental method of this kind, little could be expected 

 by way of accurate result. 



3. In nearly all the experiments (excepting those of Leslie referred to above) 

 hitherto made to determine the law of cooling, the cooling body has been placed either 

 in (so-called) still or free air, or in a vessel containing air whose pressure is different 

 from that of the atmosphere, or in a (so-called) vacuum. It has always appeared to me 

 that experiments of the kind must lead to results of a doubtful type, owing to the 

 indefinite character of the conditions under which they are made. The shape of the 

 cooling body must undoubtedly affect the direction, as the size will affect the speed, of 

 the currents of air by means of which convection is carried on ; } and, as a consequence, 

 experiments made with cooling bodies of different shapes and sizes are not comparable 

 so far as the determination of emissivity is concerned. The unsteadiness of so-called 

 still air must also affect the results, for, as will be shown later in this paper, currents of 

 air of speed so low as to make them almost imperceptible to the unaided senses, are 

 sufficient to exercise an appreciable effect on the rate of cooling of a body. It would, 

 in short, appear as if a carefully-defined unit of ' convectivity ' (to coin a word for con- 

 vection analogous to that for conduction) were required. 



4. The method adopted in the present inquiry consisted in exposing a heated body 

 to the cooling action of currents of air of different speeds, determining the temperature 

 of the body at successive intervals of time, and thereby estimating the rate of cooling 

 at given excesses of temperature at different speeds of the air current. 



The apparatus employed is represented in fig. 5, and consisted essentially of the 



* Tlieorit Analytique de la Ckalew, eh. i. sect. ii. 



t Leslie, An Experimental Enquiry into the Nature and Propagation of Heat, London, 1804, p. 279. 



X See Porter. Phil. May., xxxix. 268-279. 



