124 On the "R elation between the Changes of 



To adapt this table to the case of any other initial tempera- 

 ture than 60°, it is only necessary to remember that the dif- 

 ference of temperature {t) will be j^^th more than the table 

 gives, for every degree that T is higher, and ^^^th less for 

 every degree that T is lower. 



These results agree with remarkable precision with the 

 mean of the valuable experiments of Professor Piazzi Smyth, 

 — the difference being quite within the errors of experiment. 

 The experiments alluded to were performed on an unusually 

 large scale, with the air constantly replenished during long 

 periods, the only trustworthy mode of dealing with such 

 attenuated matter as air. 



We see, also, from inspecting the table derived from the 

 above formula, what we might otherwise have supposed to 

 be very probable, namely, that although infinite compression 

 of air would develope infinite temperature in it, yet, in ex- 

 panding it, the resulting cold onlg approaches that zero of 

 absolute cold which our formula seems spontaneously to sug- 

 gest and describe (although that idea was not designedly in- 

 troduced in forming it), and that any high degree of expan- 

 sion cannot cause the air to become colder than that, but 

 only to approximate nearer to that limit. 



We learn also what a considerable portion of the changes 

 in atmospheric temperature must be due to the barometric 

 fluctuations, — the change being one degree of temperature 

 for 0-237 inch of barometric change (at 30° barometer and 

 60° thermometer). The subject of this paper has, therefore, 

 especial importance in a meteorologic point of view, to enable 

 us to apply a proper correction to the observations of atmo- 

 spheric temperature, for that portion of it due to variations 

 in barometric pressure. 



In reference to the inventions of Professor Piazzi Smyth 

 for cooling mines or apartments in hot climates, and for ob- 

 taining a warm temperature in certain cases, we here find 

 that there is a certain amount of expansion of the air which 

 gives the maximum of cooling effect; for we have to consider, 

 in this case, not merely the amount of decrease of sensible 

 temperature, but also the quantity of matter, air, in v\hich 

 such decreased temperature exists, and the practical result 



