•208 PHILOSOPHICAL TRANSACTIONS. [aNNO 1777. 



authority in tliis inquiry, as they were made with barometers a great way distant 

 from each other, viz. near 30 miles: besides which, the geometrical heights are, 

 for the same reason, not so accurately ascertained. I have, however, ventured 

 to make what use I could of them, viz. to show tluit these too give a result on 

 the same side, though not exactly the same ; and to urge the necessity of a cer- 

 tain vicinity in those observations from which a theory is to be deduced. 



Shall I be permitted to adduce another proof, in confirmation of what has 

 been advanced ? When I first took up the consideration of measuring altitudes 

 in the atmosphere with the barometer, and had heard only of Mr. De Luc's la- 

 bours, it occurred to me that there was a much more simple method of arriving 

 at this theory, than either he or I have since pursued. It was this ; to deter- 

 mine hydrostatically the specific gravities of air and quicksilver, with a given 

 temperature and pressure ; the increase of volume, or change of gravity, with a 

 given increase of heat being supposed to be known by the experiments of Boer- 

 haave and Hawkesbee, which might be further examined by similar ones ; and 

 presuming that the geometrical ratio in the air's density, in advancing upwards 

 from the earth's surface, had been sufficiently demonstrated. For the propor- 

 tional gravity of quicksilver to air will express inversely the length of two equi- 

 ponderant columns of these Huids, that is, when the columns are taken in- 

 finitely small.* With these ideas I made the following experiment. I caused a 

 glass vessel to be blown something like a Florence flask, or rather larger ; to the 

 neck of this was adapted a brass cap with a valve opening outwards, and made to 

 screw on or off, with a male screw, by which it was fixed to an excellent pump 

 of Mr. Nairne's construction, and exhausted of its air, or at least rarefied to a 

 known degree : the vessel was then carefully weighed with a sensible balance, and 

 again after the air was re-admitted ; the difference gave the weight of the air that 

 had been exhausted. After having repeated this 2 or 3 times, the vessel was 

 exactly filled with water as far as the valve, which had been the term of capacity 

 for the air ; this was done by screwing on the cap till the superfluous water oozed 

 all out, and on inverting the vessel tliere appeared not the least sign or bubble of 

 air ; I therefore concluded, that the volume of water was precisely the same as 

 had been the volume of air, a circumstance that slK)uld be accurately attended 

 to. It was then carefully weighed, and compared with its weight when full and 

 deprived of its air. It will readily be seen, that I had then the specific gravity of 

 the two fluids, on supposition that the figure of the glass had not altered by 

 pressure during the experiment ; and this effect may be presumed to have been 



* I am not sorry (says Sir G.) to anticipate tlie reader's remark here, that this observation is not 

 new ; since I find that I have been treading the same steps with Mr. Boyle and Dr. Halley, who both 

 made use of this method ; the one with a view to determine the limits ot' die atmosphere ; and the 

 other tlie height of Snowden. — Orig. 



