Atmospherical Phenomena, 263 



ing of the barometer. It is easy to be convinced of the truth 

 of this proposition in the following manner. We observe the 

 barometer and thermometer at a particular place for some time 

 at certain hours of each day ; we determine each time, the amount 

 of change that takes place in each of these instruments, in con- 

 sequence of the irregular changes of the weather between one 

 observation and that made at the same hour on the following 

 day ; further, we distinguish the separate cases in which each of 

 these instruments rose or fell, and as we are in search of the 

 changes of temperature corresponding to the variations of the 

 pressure of the air, we determine how much the heat altered, 

 when the quicksilver rose or fell 1, 2, 3 ... lines. Observations 

 in different parts of Europe, in Iceland, on the east coast of 

 Asia, and near the Equator in South America, have confirmed to 

 me, not only the proposition announced above, viz. that a dimi- 

 nution of temperature is combined with a rise, and an increase 

 with a fall of the barometer, but have also even furnished the 

 same numerical quantities for the relation between the two 

 changes. Although the fact is thus proved in a general way, 

 yet, in reducing daily registers, we find in this point of view 

 many exceptions from the general rule. If the place lies some- 

 where between E and F, it happens not unfrequently that the 

 heat at the surface becomes greater, and that the barometer not 

 only does not sink, but actually, contrary to the rule, rises. In 

 considering this anomaly, we must not overlook the circumstance 

 that the barometer indicates the pressure of the whole atmos- 

 phere to its extreme boundaries at the point of observation, 

 whereas the thermometer indicates only the local temperature 

 of the place where it is suspended to a distance of but a few 

 feet. But the comparison spoken of above requires, that we 

 should know the mean temperature of the atmosphere to its up- 

 per limits ; and although this depends, in a general way, on the 

 indication afforded by the thermometer at the surface, yet it can 

 easily happen in particular cases, that the mean temperature of 

 the whole mass of air may diminish, while it becomes higher at 

 the surface, and vice versa. If it is difficult to find sufficient 

 grounds for these anomalies in this cause, the investigation be- 

 comes still more so, owing to the following fact. I have assumed 

 that the air is unusually heated above EF, and, in consequence 



