132 EFFECT'S of HEAT 



I have exerted myfelf to remove, by frefh experiments, every 

 doubt that has arifen on the fubjecl:. In order to guard againft 

 natural impurities, I have applied to fuch of my friends as have 

 turned their attention to chemical analylis, (a branch of the 

 fcience to which I have never attended,) to furnifh me with 

 carbonate of lime of undoubted purity. To obviate the con- 

 tamination arifing from the contact of the porcelain tubes, I de- 

 termined to confine the fubjecl: of experiment in fome fubftance 

 which had no difpofition to unite with the carbonate. I firft 

 tried charcoal, but found it very troublefome, owing to its 

 irregular abforption of water and air. 



I then turned my thoughts to the conftruction of tubes or 

 cups of platina for that purpofe. Being unable readily to pro- 

 cure proper folid vefTels of this fubftance, I made ufe of thin 

 laminated plates, formed into cups. My fir ft method was, 

 to fold the plate exactly as we do blotting-paper to form a filter 

 (Fig. 26.)', this produced a cup capable of holding the thin- 

 ner! liquid ; and being covered with a lid, formed of a fimilar 

 thin plate, bent at the edges, fo as to overlap confiderably 

 (Fig. 28.), the carbonate it contained was fecured on all fides 

 from the contact of the porcelain tube within which it was 

 placed. Another convenient device likewife occurred : I 

 wrapt a piece of the plate of platina round a cylinder, fo as 

 to form a tube, each end of which was clofed by a cover 

 like that juft described (Fig. 27. and 29). (In figure 26. and 

 27. thefe cups are reprefented upon a large fcale, and in 28. 

 and 29. nearly of their actual fize). This laft conftruction 

 had the advantage of containing eight or nine grains of car- 

 bonate, whereas the other would only hold about a grain 

 and a half. On the other hand, it was not fit to retain a 

 thin liquid ; but, in moft cafes, that circumftance was of no 

 confequence ; and I forefaw that the carbonates could not 



thus 



