The range of a liquid thermometer must not only be confined between its boil- 

 ing and freezing points, but within still more narrow limits ; for it has been 

 proved that the expansion of liquids, as they approach those temperatures at 

 which they pass into the solid or gaseous state, are subject to irregularities, 

 which render them an uncertain measure of temperature. In the choice of a 

 liquid for a thermometer we must necessarily be directed in some degree by the 

 purpose to which the instrument is applied. An instrument intended to meas- 

 ure very low temperatures may be constructed with a liquid which itself boils 

 at a low temperature ; while, on the other hand, such a liquid would be inap- 

 plicable in a thermometer designed for measuring higher degrees of heat. 

 Thermometers intended only to measure high temperatures might, on the other 

 hand, be constructed of a liquid, like certain oils, which solidifies at a consid- 

 erable temperature. For all ordinary purposes, however, that liquid will be 

 the best adapted for thermometers in which, while the freezing and boiling 

 points are separated by a great interval, that interval shall comprise the tem- 

 perature of the most ordinary objects of domestic or scientific inquiry. 



Among liquids, there is one which eminently fulfils these conditions, and 

 which, by reason of its various physical and chemical qualities, is otherwise 

 well adapted for the purposes of the thermometer. This liquid is mercury, or 

 quicksilver. Mercury boils at a higher temperature than any other liquid, ex- 

 cept certain oils ; and, on the other hand, it freezes at a lower temperature than 

 all other liquids, except some of the more volatile, such as alcohol, or ether. 

 Thus a mercurial thermometer will have a wider range than any other liquid 

 thermometer. It also is attended with this convenience, that the extent of 

 temperature included between melting ice and boiling water stands at a con- 

 siderable distance from the limits of its range. Thus it happens that nearly all 

 the temperatures which are necessary to be observed, whether for domestic 

 purposes or scientific inquiry, fall within the range of a mercurial thermometer. 

 It is attended vnth the further advantage of a higher susceptibility to the action 

 of heat, and its indications are therefore more immediate than those of other 

 liquids. Its expansibility within the extent of temperature of the phenomena 

 most commonly observed are perfectly regular, and proportional to those of sol- 

 ids and gases at the same temperatures. These properties have brought mer- 

 curial thermometers into general vise in all parts of the world. 



To render the thermometer practically useful, it is necessary that its indica- 

 tions should be steady and uniform, and capable of being compared one with 

 another at different times and places. To accomplish this, it is chiefly neces- 

 sary that the mercury which is used in different thermometers should be per- 

 fectly the same. To insure this identity, it is necessary that the mercury used 

 should be pure and free from any admixture of foreign matter. Mercury, how- 

 ever, under ordinary circumstances, is never found in this state In the mine 

 it is commonly mixed with other substances, which by chemical combination 

 render it solid, and from which it must be disengaged by the process of metal- 

 lurgy. Even when it is found in the liquid state, it is commonly mixed with 

 silver, lead, or tin — metals with which it combines with great facility. In order 

 to have it perfectly pure, it is necessary first to disengage it from the grosser 

 substances with which it may be mixed. This is easily accomplished by 

 straining it through a piece of chamois leather ; the subtle parts of the mercury 

 will pass freely through the pores by merely squeezing the leather between the 

 fingers, and the solid impurities with which it is mixed will be thus intercepted 

 and separated. 



It is still necessary, however, to disengage from the mercury other liquids 

 which may be combined with it. This is easily accomplished. Let a boiler 

 be provided, terminated in a tube at the top, which tube is conducted into a re- 



