314 A method of cooling the Air of [No. 10, new series 



a certain quantity of air will rise, on experiencing a given compres- 

 sion. On this, combined with the cost of mechanical power at the 

 place, will depend the expense and the consequent feasibility of 

 the method. Seeing that that can be accomplished by this method, 

 -which has not yet been brought about by any other, it is probable 

 that it might be adopted by the wealthy who are dying from heat, 

 although it might be very expensive ; and in hospitals, also, where 

 many subjects are concentrated together, and are more imme- 

 diately in want of the benefits of cool air — the plan might be adopt- 

 ed, although very troublesome ; but it fortunately turns out that 

 the thermotic expansion is so very great, that the machinery can 

 therefore be made very simple, and can be worked so cheaply, 

 that private persons of ordinary means may indulge in the luxury ; 

 and a house may be cooled in India for probably about the same 

 that one can be warmed in England. 



The mere fact of compression and expansion having a thermotic 

 effect on air had long been known, but no one seems ever to have 

 thought of applying it to any decidedly useful purpose, certainly 

 not this one ; and for that reason, perhaps, the exact quantity of 

 thermotic effect had never been investigated with precision ; and 

 when this idea first occurred to me in 1843. I could procure no 

 data which would enable me to calculate its practicability within 

 any moderate limits. The next year, however, I had a small appa- 

 ratus constructed for testing the matter experimentally ; and though 

 no great exactness was arrived at, still it appeared that sufficient 

 grounds were obtained to warrant the communication of the idea 

 to several friends in 1845, as a possible mode of accomplishing the 

 end in view. In 1847, I had a larger apparatus made, and in the 

 beginning of 1849, communicated an account to the Royal Society 

 of Edinburgh. 



Experiments, however, with small apparatus, are very uncertain, 

 where heat is concerned ; and in this case the results were not by 

 any means so favourable as they might have been, on account of 

 the great radiation and conduction of heat, due to the diminished 

 size of the metallic vessels, and the consequence preponderance of 

 surface to cubical contents. But in the latter end of 1849, I was 



