OF STEAM AT DIFEEEENT TEMPEEATUEES. 
207 
error, if this point were in all cases determined by a reduction taken from observations 
made at a point of temperature somewhat above that of saturation. 
For small ranges of temperatm'e, we may presume that the relation expressed by 
equation (1) will apply within small limits of error to superheated steam. Let 
be the specific volume, pressure, and temperature, respectively, of a given weight of 
steam when superheated; then we find 
(4.) 
where the constant a=74’8 nearly. Moreover, let Tj be the maximum temperature of 
saturation of the steam, and Vj its corresponding specific volume, then, from equation 
(1), we get 
(5.) 
and hence 
P, = kf+^(T.-<.) (6-) 
Now having given from the Tables of the experiments the values of taking 
Vj as the specific volume at the maximum descent of the mercury, the value of T, may 
be assumed with a near approach to accuracy from the indications in the experiments ; 
then, if the value of Pi deduced from this assumption coincide with the pressure of 
saturated steam at that temperature, the maximum temperature of saturation has been 
rightly assumed. If not, the value of Tj deduced from the value of Pj given by the 
equation may be taken for a second approximation, and so on till the value of Pj does 
coincide with the pressui’e of saturated steam deduced from the experiments of 
M. KEGItAULT. 
The following Table gives the temperature of saturation thus deduced from the pre- 
cediag experimental Tables, the two highest temperatures of superheating being in most 
cases assumed as the data for calculation. In one or two cases, where the lower of these 
two temperatures is manifestly within the limits of imperfect expansion, the reduction 
from the higher temperature alone has been retained. 
The point thus determined (strictly speaking) is the temperature of maximum satura- 
tion of the steam, assuming that at this point it retains all its watery vapour in an 
aeriform state. This ideal point of temperature, it must be observed, will always be a 
trifle higher than the actual point. As the coetficient a is always small as compared 
with V,, the value of Pj, in equation (6.), is not much affected by any possible variations 
of value which may be given to this coetficient. 
