VOL. LXVII.] PHILOSOPHICAL TRANSACTIONS. 209 



the most sensible, when the vessel was filled with water, the pressure at that 

 time being from within. To assure myself of this, I let in a small quantity of 

 air, which formed a bubble of about J- of an inch in diameter, and on immerging 

 the glass in another vessel of water, by which the pressure within was counter- 

 poised by a pressure without, the bubble seemed to contract itself by a quantity, 

 as I found afterwards, equal to about 2 grains in weight, or t-sVo- of the whole 

 contents. I therefore concluded, that this correction was hardly worth taking 

 notice of, and still less the effect from external pressure when the glass was ex- 

 hausted. At every operation the height of the barometer and thermometer 

 (placed close to the vessel when the air was weighed) was noted down, together 

 with the height of the pump-gage, which, compared with the barometer in the 

 room, showed the quantity exhausted. The result of the experiment was as 

 follows, the barometer in the room standing at 2().27 inches, and the heat of 

 the room 53°. 



Grains. 



The bottle empty or exhausted till the gage stood at 29.15 inches weighed (determined 



from -i diiFerent trials, and die balance turning with JL of a grain) 2657.40 



Increase of weight when filled with air, from i trials certain to .j!^ of a grain + l6. 13 



Bottle filled with water, whose heat was 51" , 16.120.00 



Weight of the water, exclusive of the bottle 13562.6o 



But the bottle was exhausted only in the proportion of 20. 1 5 inches to 29.27 



inches ; therefore if a perfect vacuum could have been made, the difference of 



weight would have been l6.22 grains instead of 10.13 grains. Again, the 



water was colder than the air by 2° ; the one being 53°, and the other only 5 1° : 



now water, from former experiments, I find to expand about -nrhro with 2° of 



heat ; therefore, if the water had been of the same temperature with the air that 



was examined, the weight of an equal volume would have been only 13558.5 



grains; and lastly, 13558.5 divided by 1 6.22 gives 836,* and by so much is 



water heavier than air in these circumstances. 



By former experiments I find the specific gravity of the quicksilver of my barometers, 



compared with rain-water in 68" of heat, as 13.606 to 1 



And 68" — 53" = 15°, correct tlierefore for 15° of expansion of quicksilver + .018 



Correct for 15° of expansion of air — .03 1 



True specific gravity of quicksilver, with 53° of heat 1 3. 59+ 



Which multiplied by the specific gravity of air X .836 



Gives for the comparative gravity of quicksilver and air, when the barometer is at 



29.27, and the thermometer 53° 1 136-1.6 



* Hawkesbee's experiments made the air 850 lighter than water, the barometer being at ?g.7 ; 

 and Dr. Halley supposed it abort SCO. Mr. Cavendish, in weighing 50 grains of air, when the baro- 

 meter was at 2i)-J, and the thermometer at 50°, concluded the specific gravity of air to be about 800 

 also. Now my experiment, reduced to the same circumstances with his, would give 817 for this 

 gravity, no great difference in an affair of such delicacy. — Orig. 



VOL. XIV. E E 



