Datiiell on a new Hygrometer. 323 



From these it will be observed, that, at the period of the ob- 

 servations, the temperature upon the top of St. Paul's, con- 

 trary to the generally received opinion, was, upon the average, 

 1°.2 higher than at the bottom. I no sooner perceived this, 

 than I placed a register thermometer there, and compared it 

 with one placed as nearly as possible under similar circum- 

 stances below. The temperature of the first was often 6° higher 

 than that of the second, and the mean of all the observations 

 was 2°.9 warmer. The mean point of condensation was 0°.9 

 lower above than below, and the mean difference between the 

 temperature of the air and that of the vapour 2°. greater. The 

 peculiarities of the season when these experiments were tried, 

 must be taken into the account. Fogs and hoar-frost were very 

 prevalent. The surface of the earth was cooled by nightly ra- 

 diation, which lowered the temperature of the contiguous air. 

 This influence could extend but a small way, and ice of con- 

 siderable thickness was often formed on the earth, while the 

 thermometer never once reached the freezing point at the upper 

 station. We herein have the history of a hoar-frost, and the 

 reason of its short continuance. It is only when it has its 

 origin in the upper regions that a frost can be of long continu- 

 ance ; but this is one of the points of theory that I have 

 reserved for my next paper, and I shall therefore say no more 

 upon it at present. The few experiments that I have had it in 

 my power to make at the summit of St. Paul's, sufficiently 

 prove, that a regular series of meteorological observations upon 

 that elevation, compared with one below, would be interesting 

 and instructive. 



Tlie last application of the Hygrometer, which I shall point 

 out, is perhaps of superior importance to any of those which 

 we have been considering. I mean the correction which it af- 

 fords to barometrical measurements. The principle upon which 

 the Barometer is at present applied to the determination of 

 heights, is the gradation of the density of the atmosphere, con- 

 sidered as a homogeneous fluid of uniform composition. The 

 only correction at present applied is an allowance for the dis- 

 turbing influence of heat by the expansion of the air, and con- 



