o/ Air' and Mercury- Thermometers. 323 



which represents 100°, we deduce 0°-0618, or about y»^th as the 

 fraction of a degree which the air-thermometer is in advance of 

 true temperature at 50°. 



At 200°, by a similar computation, it is |th of a degree behind 

 the true temperature. 



§ 18. It will be remarked that the differences are of contrary 

 signs to the differences between the mercmy- and air-thermome- 

 ters, and that between 0° and 200° ; a reduction of ^th in the 

 correction of the mercury-thermometers derived from the formula 

 [Appendix, III.] would compensate almost exactly for this slight 

 aberration from uniform expansion. 



Appendix. — I. 



This expressed by symbols is \/^ [=AE, fig. 4], in which e is 



the tension and t the temperature [=CE], by air-thermometer 

 reckoned from [C] the zero of gaseous tension. 



The ranging of the points in a straight line or gradient is shown 

 by the quotient of the difference of ordinates by difference of abscissae 



being constant: thus [g^] = _^^LZL_ = A = a con- 



VrrVT 



stant ratio = cotangent of the inclination of the gradient to the axis 

 of temperature = — •, fig. 4 . 



The value of ^o when e=0 is thus to=ti—h^/ J=^= the tem- 

 perature of the zero of tension of aqueous vapour [=CG]. 



These constants, h and g, may thus be obtained from two observa- 

 tions of tension, and the general equation for the tension in terms of 

 the temperature is 



Assuming AS (fig. 2) to be the true gradient of steam, we have 



logL2-1962446r'"^'^'=^°^^^^"'"^^' 



1 { r.«r,oif,^i v English measures. 

 logA=2-7281741 J '=' 



This gradient coincides nearly with the observations of Dr. lire 

 and the American Institute at high pressures, but it is too high for 

 those of the French Academy ; also for those of M. Regnault above 

 100°. These last were taken by a process different from that em- 

 ployed for the observations below 100°. On examining them mi- 

 nutely I find that they are not continuous ; there is a dislocation at 



