1818.] the Imperfection of the Barometer. 199 



this object, I applied for the assistance of a friend at Lyme, in 

 Dorsetshire,* who was much in the habit of making meteoro- 

 logical observations, and on whose accuracy I could place full 

 confidence. The register kept at the Royal Society gave me a 

 second source of correct information ; and my own observations, 

 made at Knotshole, near Liverpool, where I then resided, 

 afforded a third station ; by means of which the state of the 

 atmosphere might be compared in three points of the kingdom, 

 at a considerable distance from each other, and adjacent 

 respectively to the southern, eastern, and western oceans. The 

 result of my inquiry was not very satisfactory as to the particular 

 object I had in view; but it led me to some ideas respecting the 

 use of the barometer, which, I believe, are a little different 

 from those generally entertained. 



It is well known that there is a correspondence between the 

 state of barometers that are situated at considerable distances 

 from each other ; that when the mercury is high or low in 

 London, it is high or low in Paris ; f but it does not appear that 

 any very minute examination of the subject has been made. 

 We have been satisfied with ascertaining the fact of the general 

 agreement of the instruments ; and if any discrepancies have 

 been noticed, they have been regarded as arising from some 

 accidental circumstance ; and no inquiiy seems to have been 

 made how far they depended upon an absolute difference in the 

 weight of the atmosphere, or how far upon any defect in the 

 nature or construction of the apparatus. Upon comparing 

 together the registers of the three barometers kept respectively 

 in London, Lyme, and Liverpool, a general correspondence was 

 found between them ; but at the same time the variations were 

 very numerous, and of a kind for which I found it very difficult to 

 account. PL LXXVIII. Fig. No. 1, contains aview of the heights 

 of the three barometers during the months of September and 

 October in the year 1811, taken nearly at the same hour in the 

 morning of each day.'j These two months were selected, 

 because I was able to procure a greater number of accurate 

 observations for them than for any other period, and, in some 

 degree also, in consequence of the state of the weather, which, 

 during the first part of September, was unusually fine and 

 settled, with steady east winds ; while the month of October was 

 equally remarkable for its variable character, with the wind 

 chiefly blowing in the opposite direction. In examining the 



* Geo. Holland, Esq. 



+ An interesting account of (lie comparative state of the barometer in London, 

 Paris, and Geneva, during the space of a year, from the autumnal equinox in 

 Ib06 to the autumnal equinox in 1807, was published by M. Pictet in the Bib. 

 Brit, for Jan. 1811, of which Dr. Thomson has given a translation in vol. ii. of 

 the Annals. 



:f The observations at Lyme and Liverpool were always made at nine o'clock; 

 the time of the observations at the apartments of the Royal Society varies a little, 

 but is generally at or between the hours of eight and nine. 



