364 Dcfcripi'mi of a pertailg Haromiter. 



meiit of crcat Iieights arife rather from the unequal denfitiea 

 of the itrata af the atmoiphere, and the attraclion of the njer- 

 curv in the tube, than from defeats in the fubdivilion. A ba- 

 rometer will always be pretty good, if the mercury it contains 

 is well puriiled and dry; and if it has been boiled in the 

 tube ; if the tube is at kaft |ths of a line in diameter ; if, by a 

 proper conltrucVi )n, the air is prevented from introducing it- 

 felf into it, notv.ithllanding the agitation it is expofed to in 

 the courfe of a long journey ; and, in particidar, if the inter- 

 yal between the furface of the mercur\' contained in the tube 

 autl that in the cup can be exaflly meafured. 



This barometer difters very little from that of Deluc ; but 

 it is much fimplor in the conftru(!:lion, and lefs expenfive. 

 The cup which I have fubftiluted for the tube, covered with 

 leather, ferves to retain a little mercury in the refervoir, that 

 no vacuum mav remain between the itop-cock and the bend- 

 ino", when the barometer is inclined for the piirpofe of being 

 tranfportcd. As it is not deep, and has a cylindric form, it 

 may be eafily cleaned in cafe the metal fliould become oxi- 

 dated. Deluc acknowledges that his barometer is inconve- 

 nient for daily obftrvation. The one here defcribed has not 

 the fame fault. It is both portable and fit for the clofet. I 

 have already faid that, neglecting all corre6lion, the error re- 

 fultiiig from the difference of the levels would amount only to 

 ^'-d—^a difference which may be negletSled in making meteo- 

 iological obfervations. 



Oiie of the greated advantages of this inftrument is, that it 

 ■jnay be rendered comparative, by choofing tubes of about two 

 Jines diameter, and cylinders of an inch. It is of no im- 

 .portance whether a cup-barometer maintains itfelf a little 

 higher or a little lower than a fyphou oiie, if the depreffion is 

 .eontiant, and its progrefs regular. 1 have carried both baro- 

 n5,eters on mountains more than 1800 toifes in height. 



I made a great many experiments on this lubjeft on 

 Mount Ventoux. Mine I confefs always maintained itfelf a 

 line lower; but was this depreffion of a line a fault of the 

 cup-barometer, or did the fault confiil in, the greater eleva- 

 .tiou of the other? I do not know whether tiie folution of 

 this problem wouki be favourable to Deluc's barometer, or 

 to mine. 



If the barometric obfervations do not require too much 

 time and preparation ; and if a fimple, Itrong, and cheap 

 conftrut^tion fupply the place of complex indrimients liable 

 to be deranged ; I can aflert that all thefe advantages are 

 united in this^baromeler, by means of which correal obferva- 

 tions may be made, whick will fupply our fcicnlific inllitu- 



tions 



