OF ARTS AND SCIENCES. 77 



on the air thermometer decreases from 0° to about 30° or 35°, after 

 which it increases, seems to be an entirely unique fact in nature, seeing 

 that there is apparently no other substance hitherto experimented upon 

 whose specific heat decreases on rise of temperature without change of 

 state. From a thermodynamic point of view, however, it is of the 

 same nature as the decrease of specific heat which takes place after 

 the vaporization of a liquid. 



The close agreement of my result at 15°.7 C. with the old result of 

 Joule, after approximately reducing his to the air thermometer and 

 latitude of Baltimore, and correcting the specific heat of copper, is 

 very satisfactory to us both, as the difference is not greater than 1 in 

 400, and is probably less. 



I hope at some future time to make a comparison with Joule's 

 thermometers, when the difference can be accurately stated. 



II. — THERMOMETRY. 



{a.) General View. 



TRe science of thermometry, as ordinarily studied, is based upon 

 the changes produced in bodies by heat. Among these we may men- 

 tion change in volume, pressure, state of aggregation, dissociation, 

 amount and color of light reflected, transmitted, or emitted, hardness, 

 pyro-electric and thermo-electric properties, electric conductivity or 

 specific induction capacity, magnetic properties, thermo-dynamic prop- 

 erties, &c. ; and on each of these may be based a system of ther- 

 mometry, each one of which is perfect in itself, but which differs from 

 all the others widely. Indeed, each method may be applied to nearly 

 all the bodies in nature, and hundreds or thousands of thermometric 

 scales may be produced, which may be made to agree at two fixed 

 points, such as the freezing and boiling points of water, but which 

 will in general differ at nearly, if not all, other points. 



But from the way in which the science has advanced, it has come 

 to pass that all methods of thermometry in general use to the present 

 time have been reduced to two or three, based respectively on the 

 apparent expansion of mercury in glass and on the absolute expan- 

 sion of some gas, and more lately on the second law of thermo- 

 dynamics. 



Each of these systems is perfectly correct in itself, and we have no 

 right to designate either of them as incorrect. We must decide 

 a priori on some system, and then express all our results in that 

 system : the accuracy of science demands that there should be no 



