Notes on Barometer Construction. 47 



very different to what we call boiling. Now, oil of vitriol, 

 methylic alcohol, and mercury — most probably on account 

 of the absence of dissolved gases — are each converted 

 into vapour with more or less tendency to sudden bursts 

 and "bumpings," as they are called, and in these cases 

 the distillates are liable to contamination with portions 

 of the fluid, scattered and thrown over rather than distilled ; 

 and some kind of artifice is requisite in all such cases for 

 obviating this source of an imperfect resujt. Many j^ears 

 ago a French chemist (M. Yiolette) recommended the use of 

 super-heated steam for the distillation of mercury — a pro- 

 mising suggestion enough ; but a purification completely 

 satisfactory may be effected by simpler means. Three or 

 four circular discs of iron wire gauze are allowed to float on 

 the mercury in the retort, covering its whole surface ; or, 

 what is better, a layer of three-quarters of an inch of small 

 cut or wrought iron brads are allowed to float on the metal ; 

 either of these forms a mechanical barrier, holding back the 

 mechanically dispersed fluid mercury, but allowing suffi- 

 ciently free escape for the mercurial vapour. For the reason 

 already given a very small stream only of cold water, run- 

 ning over a cloth laid over the exit-tube of the iron retort, 

 is requisite for re-condensation of the mercury. The lower 

 end of the exit-tube is also bound round with a few folds 

 of calico, which, projecting beyond it^ form a tubular conduit 

 sufficient for confining and conducting the condensed mercury 

 into a pan of water, and at the same time sufficiently per- 

 vious to the atmospheric air to prevent the water in the col- 

 lecting pan being drawn up into the retort as a result of 

 condensation of mercurial vapour at the end of the operation. 

 I believe that a more extensive acquaintance with the 

 efficacy of this simple method of distillation would cause 

 its employment in preference to the several methods of 

 chemical treatment. 



A few observations on boiling out and other modes of 

 filling glass tubes with mercury may now be added. Boil- 

 ing out means raising the mercury to the temperature at 

 which it freely forms metallic vapour, and so expelling the 

 atmospheric air from the tube; it also means raising the 

 mercury to a temperature at which its oxidation takes 

 place when in contact with atmospheric air. The warm 

 mercury is added in small doses to the inverted tube, and 

 the boiling is brought about by heating the tube at a point 



