490 Mr. J. P. Joule and Prof. Thomson on the Thermal Effects 



The numbers in the last column of any one of these tables 

 show, by their disci'epancies, how much uncertainty there must 

 be in the results on account of purely accidental circumstances. 



The following table is arranged, with double argument of tem- 

 perature and of quantity of air passing per second, to show a 

 comparison of the means of the different series (Scries 3 being 

 divided into two, one consisting of the first five experiments, and 

 the other of the remaining two). 



Table of Mean Values of -, — =:r in different series of experiments. 

 losP 



The general increase of the numbers from left to right in this 

 table shows that very much of the cooling effect must be lost on 

 account of the insufficiency of the current of air. This loss 

 might possibly be diminished by improving the thermal insula- 

 tion of the current in and after the rapids ; but it appears pro- 

 bable that it could be reduced sufficiently to admit of satisfactory 

 observations being made, only by using a much more copious 

 current of air than could be obtained with the apparatus hitherto 

 employed. 



The decrease of the numbers from the upper to the lower 

 spaces, especially in the one complete vertical column (that under 

 the argument 5*6), shows that the cooling effect is less to a 

 remarkable degree for the higher than for the lower temperatures. 

 Even from 41° to 65° F. the diminution is most sensible ; and 

 at 171° the coohng effect appears to be only about half as much 

 as at 41°. 



The best results for the different temperatures are probably 

 those shown under the arguments 8'4 and 11*2, being those 

 obtained from the most copious currents ; but it is probable that 



they all fall considerably short of the true values of ^ — p for 



the actual temperatures ; and we may consider it as perfectly esta- 

 blished by the experiments described above, that there is a final 

 cooling effect produced brj air rushing through a small aperture at any 

 temperature up to 170° F,, and that the amount of this cooling effect 



