ceiver, placed beyond the influence of the fire, so as to be capable of recon- 

 verting the vapor of mercury into liquid. Let the impure mercury be placed in 

 this close boiler on a fire. The fact that mercury boils at a lower temperature 

 than any other metal, will cause it to be converted into vapor, while the other 

 metals with which it is mixed continue in the liquid or solid state. The mer- 

 cury will thus pass over in vapor through the pipe from the top of the boiler 

 into the cooler, where it will be restored to the liquid state, and will be col- 

 lected free of admixture with other metals. This process, which is called dis- 

 tillation, will be more fully described hereafter. If the mercury happen to hold 

 in combination any liquid which boils at a lower temperature than the mercury 

 itself, such a liquid may be dismissed by raising the mercury in the boiler to a 

 temperature below its own boiling point. The liquids combined with it will 

 then pass over in vapor, and will be collected in trie cooler separate from the 

 mercury. 



Having now obtained pure mercury, unalloyed by admixture with any other 

 substance, the next object is to contrive a means of rendering its dilatations and 

 contractions observable. For this purpose, let a glass tube, of very small bore, 

 be obtained by the ordinary process of glass-blowing ; let a spherical bulb be 

 blown at one end of it, of a magnitude very considerable compared with the 

 bore of the tube. As the tube must be of that extremely small bore which is 

 called capillary, the bulb, though not of great magnitude, may still bear a very 

 considerable proportion to it. When the bulb is filled, a very slight change in 

 the volume of the mercury will cause a considerable rise or fall in the tube ; 

 because the bulb not considerably altering its dimensions, an increase of vol- 

 ume in the mercury must necessarily find room by forcing the column upward 

 in the tube ; and a diminution of volume, for a like reason, will cause the col- 

 umn in the tube to fall. If a portion of the bore of a tube, measuring the eighth 

 of an inch in length, contain the 1000th part of the whole quantity of mercury 

 in the apparatus, then an expansion, amounting to one part in 1000, will cause 

 the column of mercury to rise in the tube the eighth of an inch, a space which 

 is easily observable ; and if the bore of the tube be everywhere uniform, 

 every eighth of an inch which the column of mercury rises or falls will 

 correspond to an equal increase in the volume of mercury. The tube 

 and bulb, thus constructed, are attached to a divided scale, by which the rise 

 or fall of the column of mercury in the tube may be accurately measured and 

 observed. 



If the scale by which the variations of a mercurial column are measured be 

 divided into equal parts, it is obvious that the bore of the tube should be uni- 

 form, for otherwise equal divisions of the scale would not correspond to equal 

 dilatations or contractions of the mercury. If one part of the bore were larger 

 than another, a division at that part would correspond to a greater change in 

 the volume of the mercury than a division at another part where the bore is 

 narrower. As it is a matter of convenience that the divisions on the scale 

 should be equal, it is obviously essential that the bore of the tube should be 

 either accurately or very nearly uniform. There is a very simple and effectual 

 method of ascertaining whether the bore of a tube fulfil this condition. Before 

 the bulb is blown on the tube, let a drop of mercury be introduced into its bore 

 so small as to occupy a space in the bore not exceeding a quarter of an inch, 

 or even less. Let this mercury be gradually moved through the tube from end 

 to end, causing it to rest at different points by holding the tube horizontally, 

 and let the space which it occupies in the tube at different places be measured 

 by some accurate measure. If the mercury occupies the same length of the 

 tube in every part of its bore, it is evident that the bore will be everywhere 

 uniform ; but if it occupies a less extent of the bore in one place than in an- 



