678 PHILOSOPHICAL TRANSACTIONS. [anNO 1790. 



portions being put in by means of a well known instrument, which is composed 

 of a small dish terminating in a tube drawn to a fine point: the top of the dish 

 being covered with the thumb, the liquor in it is prevented from running out 

 through the tube by the pressure of the atmosphere, but instantly begins to issue 

 by drops, or a very small stream, on raising the thumb. Water being thus in- 

 troduced into the phial, till it exactly counterpoised the weight, which, having 

 been previously computed^ was put into the opposite scale, the phial was shaken, 

 and then well stopped with its glass stopple, over which leather was tied very 

 tight, to prevent evaporation. No mixture was used till it had remained in the 

 phial at least a month, for the full penetration to have taken place ; and it was 

 always well shaken before it was poured out to have its specific gravity tried. 



2. There are 2 common methods of taking the specific gravity of fluids; one 

 by finding the weight which a solid body loses by being immersed in them; 

 the other by filling a convenient vessel with them, and ascertaining the increase 

 of weight it acquires. In both cases a standard must have been previously taken, 

 which is usually distilled water; namely, in the first method by finding the 

 weight lost by the solid body in the water, and in the 2d method, , the weight of 

 the vessel filled with water. The latter was preferred for the following reasons. 

 When a ball of glass, which is the properest kind of solid body, is weighed in 

 any spirituous or watery fluid, the adhesion of the fluid occasions some inac- 

 curacy, and renders the balance comparatively sluggish. To what degree this 

 effect proceeds is uncertain ; but from some experiments made by Mr. Gilpin, 

 with that view, it appears to be very sensible. Also, in this method a large sur- 

 face must be exposed to the air during the operation of weighing, which, especially 

 in the higher temperatures, would give occasion to such an evaporation as to 

 alter essentially the strength of the mixture. It seemed also, as if the tempera- 

 ture of the fluid under trial could be determined more exactly in the method of 

 filling a vessel, than in the other: for the fluid cannot well be stirred while the 

 ball to be weighed remains immersed in it; and as some time must necessarily be 

 spent in the weighing, the change of heat which takes place during that period 

 will be unequal through the mass, and may occasion a sensible error. It is true, 

 on the other hand, that in the method of filling a vessel, the temperature could 

 not be ascertained with the utmost precision, because the neck of the vessel 

 employed, containing about ]0 grains, was filled up to the mark with spirit not 

 exactly of the same temperature, as will be explained presently; but this error, 

 it is supposed, would by no means equal the^ other, and the utmost quantity of 

 it may be estimated very nearly. Finally, it was much easier to bring the fluid 

 to any given temperature when it was in a vessel to be weighed, than when it was 

 to have a solid body weighed in it; because in the former case the quantity was 

 smaller, and the vessel containing it more manageable, being readily heated with 

 the hand or warm water, and cooled with cold water: the very circumstance, that 



