24« IMPROVED BAROMETERS. 



the beft way is to determine it by obfervation, viz. by finding 

 how much of the rod exprefles fome larger variation in the 

 common barometer, and then taking out the rod and gradu- 

 ating, &c. or rather by taking the mean of a great number of 

 obfervations. 



Should this be found on a fufficient trial to anfwer the pur- 

 pofe, I can point out a method of making it ferve alfo as a 

 thermometer. 

 This barometer Before I conclude, I fliall juft mention the idea which led 

 refleafon on an- me to ^ s conftru&ion, which though the greateft increafe of 

 other conftruc- fcale it gives is 14:1, may poiTibly in the end be found to 

 tion - anfwer better. 



Water poured on Jt ftruck me that if we were to pour in water into a long 

 face of afyphon tu ^ e ^ F, connected with the tube of the common barometer 

 mercurial baro- as in Fig. 2, or to take out water fo as at any time to bring 

 "mete 8 r which 3 " the mercui 7 to the fixed P " 1 * N, the altitude of water either 

 may be made to poured in or taken out would give us a meafure for the varia- 



fljw variations t j ons f the barometer fourteen times larger than in thecom- 

 14 times as long . 1,1 



as the mercurial mon one > nnce water is fourteen times lighter than mercury, 

 change. This j s evident from the common hydroftatical principles, as 



the altitude of water poured in is to be a counterpoife to the 

 fum of the fall of mercury in A B, and rife in C D, and e con- 

 tra, the altitude of water taken put is to be a counterpoife to 

 the fum of the rife of mercury in A B, and fall in C D. 



If the tube DF therefore were graduated, it would give a 

 fcale 42 inches long, and the rife of water would indicate 

 diminution in the weight of the atmofphere and v. v. thence the 

 fcale to be inverted. To facilitate the pouring out and in of 

 water, let there be a bucket H moveable up and down by 

 rackwork, or otherwife, from D to F; to this let there be at- 

 tached the fyphon S. 



By railing the bucket until the level of the water in it be 

 higher than the level of that in the tube, we pour in water at 

 pleafure, and e contra, by bringing down the bucket we take 

 out water at. pleafure. 



This inftrument will have all the advantages of a water ba- 

 rometer without its unmanageable length ; nor do I fee any 

 pbjerftion to it, except perhaps that its appendages are too 

 bulky *. 



V. A Summary 



* The principal objection to the ingenious contrivances in the 



prefent communication are, that the refults originally depend on 



the 



