1871.] Heat, and Force, Sy 



thinking that this experiment is wholly inconclusive. First, 

 is it not exceedingly strange that when the engine did no 

 work the battery calorimeter absorbed only 13,888 calories, 

 but when the former raised 131*24 kilogramme-metres, the 

 latter should register 15,427, thus showing that when the' 

 engine did no work, it (the engine) exercised a much greater 

 comparative resistance and absorbed much more heat than 

 when it did the work ? Then, can any serious conclusion be 

 built on a difference of 300 units out of 18,682 ? Is not this 

 difference quite within the limits of accidental error? 

 Indeed, the difference is much less than differences shown in 

 other experiments of M. Favre where he had no engine to do 

 any work. But there is a very much more serious objection 

 than either of these. Supposing the numbers to be strictly 

 reliable, is there not a much simpler explanation of the phe- 

 nomenon ? M. Favre does not tell us how the magnetic 

 engine worked, but doubtless it worked as most of such 

 engines do (chiefly at least) by pulling iron keepers to the 

 electro-magnets. Now, by this action the iron is expanded, 

 and this pulling or expanding action, as we have shown, 

 usually produces cold, and hence the disappearance of 300 

 units of heat. If the engine had worked by pushing instead 

 of pulling, that is, by repulsion instead of attraction, should 

 we not have had an increase of heat instead of a decrease ? 

 We have every reason to conclude that we should. Un- 

 fortunately the apparatus required for repeating these experi- 

 ments is so very costly and delicate that very few persons 

 are in a position to repeat them, and M. Favre himself has 

 either never repeated them, or if he has, as he seems to have 

 done, he has never given us the full results. This one single 

 experiment is the only one of the kind of which he has published 

 the result. In the accounts of his later experiments he has 

 never published the number of the calories evolved in both 

 the battery calorimeter and the engine calorimeter, but only 

 the former ; and a calculation of what the latter ought to be, 

 but not what they actually were. 



4. And to set against this single experiment of M. Favre, 

 we have numerous experiments of M. Soret, in which he finds 

 results totally discordant with those of M. Favre. In the 

 " Comptes Rendus," xlv., 301, 380, M. Soret gives us the 

 result of his experiments. He placed an electro-dynamic 

 engine in a calorimeter to ascertain the effects of its 

 working. Unfortunately, he gives us very few details, 

 but he says (as we should expect from what we have just 

 said) that the results were very discordant with each other. 

 When using a brass calorimeter he found that the effect of 



