They are not, however, numerated commencing from either of those fixed 

 points of temperature, but the numeration commences at the thirty-second di- 

 vision below the freezing point, so that the freezing point is 32° and the boil- 

 ing point 212°. The origin of this circumstance will be stated hereafter. The 

 centigrade thermometer, used in France, has the intervals between the fixed 

 terms divided into 100 equal parts called degrees, the numeration commencing 

 at the freezing point. The thermometer of Reaumur, generally used in other 

 parts of Europe, has the intervals divided into 80°, the numeration commencing 

 likewise at the freezing point. In all thermometers, the degrees below that at 

 which the numeration commences upward are called negative, and are marked 

 by the sign — prefixed to the number. Thus, — 10° means 10° below that de- 

 gree at which the numeration upward commences. 



On the slightest consideration it will be perceived that however thermome- 

 ters may vary in the intervals between the freezing and boiling points, they 

 must, if constructed in the manner just described, agree in their indications of 

 temperature. If two thermometers, having different intervals between these 

 points, be immersed in melting ice, they will both stand at the freezing point. 

 If they then be both transferred into the water at a temperature exactly mid- 

 way between that and the temperature of boiling water, the mercury, expand- 

 ing in the same proportion in both, will dilate by exactly half that quantity which 

 it would dilate were it exposed to the temperature of boiling water ; conse- 

 quently it will stand at the middle point exactly between the fixed terms of 

 the scale, and, consequently, upon Fahrenheit's scale, it will indicate the 

 temperature of 122°, being 90° above the freezing point, and 90° below the 

 boiling point. In like manner, if the thermometer were immersed in water 

 having a temperature exceeding the temperature of melting ice by one third 

 of the excess of the temperature of boiling water above that of melting ice, 

 it is evident that the mercury will rise in both through one third of the in- 

 tervals between the fixed terms, and, consequently, would ascend through 

 a space equal to 60° of Fahrenheit above the freezing point. It would, there- 

 fore, stand in both at the temperature of 92°. This reasoning may easily be 

 generalized ; and it will be sufficiently apparent that the indications of differ- 

 ent thermometers will be the same, whatever be the length of the interval 

 between the fixed terms of their scales. 



These arrangements being made, it will be perceived that all thermometers 

 thus constructed, however different they may be in size, in the capacity of their 

 bulbs, or in other circumstances, will always be comparable with each other. 

 Experiments performed in different parts of the world may, therefore, be com- 

 municated from place to place, and repeated, with the certainty of an exact 

 correspondence ; and all the advantages arising from multiplied experience will 

 thus be obtained. 



Various other liquids besides mercury have been employed in the construc- 

 tion of thermometers ; but the several conditions for the attainment of accuracy 

 which have been explained in reference to the mercurial thermometers are, for 

 the most part, generally applicable to all liquid thermometers whatever. Alco- 

 hol, or spirits of wine, is a liquid not uncommonly used for thermometers. Its 

 inconvenience, however, for ordinary purposes, is, that it boils at a tempera- 

 ture below that of boiling water ; and, consequently, it will not admit of a scale 

 so high as this temperature. By adopting the precaution of excluding the air 

 from the tube by the method already explained in the mercurial thermometers, 

 the spirits of wine may, however, be made to indicate much higher tempera- 

 tures than is commonly supposed. They may be raised to the temperature of 

 boiling water, or even above it. If the air be perfectly excluded from the tube 

 when trie temperature is raised above the boiling point of alcohol, the upper 



