MECHANICS AND USEFUL ARTS. 



63 



speculations throw considerable doubt on the accuracy of the above formula 

 when applied to steam and other condensable vapors. Several years ago, Dr. 

 Joule and Professor William Thomson announced, as the result of applying 

 the new dynamical theory of heat to the law of Carnot, that, for tempera- 

 tures above two hundred and twelve degrees Fahrenheit, there is a very con- 

 siderable deviation from the gaseous laws in the case of steam. Later, in 

 18-55, Professor Maquorn Rankine has given a new theoretical formula for the 

 density of steam, independent of Gay-Lussac's law, and confirmatory of 

 Professor Thomson's surmise. But as yet these speculations need the evi- 

 dence and verification of direct experiment. 



The density of steam is ascertained by vaporizing a known weight of water 

 in a glass globe of known capacity, and noting the exact temperature at 

 which the whole of the water becomes converted into steam. From these 

 three elements volume, weight, and temperature the specific gravity is 

 known. But in pursuing this method, these two difficulties must be over- 

 come : First, the pressure of the steam renders it necessary that the glass 

 globe should be heated in a strong, and consequently opaque, vessel; second, 

 as steam rapidly expands in volume for any increase of temperature, beyond 

 the temperature of saturation, it would, in any case, be impossible to decide 

 by the eye the temperature at which the whole of the water became vapor- 

 ized. The temperature of saturation, or temperature at which the whole of 

 the moisture is converted into steam, while no part of the steam is super- 

 heated, must be determined with the utmost accuracy, or the results are of 

 no value. 



The difficulties thus resolve themselves into finding some other test of suf- 

 ficient accuracy and delicacy to determine the point of saturation. This has 

 been overcome by what may be termed the saturation gauge; and it is in this 

 that the novelty of the present experiments consists. 



To illustrate the principles of the saturation gauge, suppose two globes, A 

 and B (Fig. 1), connected by a bent tube containing mercury at a b, and 

 placed in a bath in which they can be raised to any desired temperature. 

 Suppose a Torricellian vacuum to have been created in each globe, and 

 twenty grains of water to have been added to A, and thirty or forty grains 

 to B. Now, suppose the temperature to be slowly and uniformly raised 

 around these globes; the water in 

 each will go on evaporating at each 

 temperature, being filled with steam 

 of a density corresponding to that 

 temperature, and the density being 

 greater as the temperature increases. 

 At last a point will be reached at 

 which the whole of the water in globe 

 A will be converted into steam ; and 

 at this point the mercury column will 

 rise at a, and sink at 6. This is the 

 saturation test, and the cause of its 

 action will be easily seen. So long 

 as vaporization went on in both A and 

 B, and the temperature was main- 

 tained uniform, each globe would contain steam of the same pressure, and 

 the columns of mercury, a and b, would remain at the same level. But so 

 soon as the water in A had vaporized, and the steam began to superheat, the 



