and the J\lode of its Commiinication. 39 



above variation ; when I found the results to be as fol- 

 lows : 



Time of cooling 



the 10 degrees 



in question. 



The instrument No. I, naked metal, . . . . 55^ min. 

 The instrument No. 2, covered vi\\h. two coatings of glue, 37^ " 



Finding that two transparent coatings of glue facili- 

 tated the cooling of this instrument even more than 

 one coating, I washed off all the glue with warm water ; 

 then, making the instrument as clean and bright as pos- 

 sible, I covered its sides with a coating of very fine, 

 transparent, and colourless spirit varnish ; and, after 

 this coating of varnish had become quite dry and hard, 

 I repeated the experiment above mentioned ; and, find- 

 ing that this covering, like that of glue, expedited the 

 cooling of the instrument, I first added a second coat- 

 ing of the varnish, and repeated the experiment again, 

 and then added two coatings more, making four in all. 

 Finding that the cooling of the instrument was more 

 and more rapid; as the thickness of the varnish was 

 increased, I now added four coatings more, making 

 eight coatings in the whole, giving time for each new 

 coating to dry thoroughly before the next was applied ; 

 but I found, on repeating the experiment with this 

 thick covering of varnish, that I had passed the limit 

 of thickness which produced the greatest effect. 



In order that the result of these experiments with 

 coatings of different thicknesses of spirit varnish may 

 be seen at one view, I shall here place them all to- 

 gether ; and I shall place by the side of each the result 

 of the standard experiment, which was made at the same 

 time with the instrument No. I, the sides which were 

 naked. 



