Specific Gravities. 317 



434. It w,ill be recollected that the specific gravity of a solid 

 body is determined by dividing its absolute weight by the loss 

 sustained on its being weighed in pure water, and the specific 

 gravity of a fluid, whether liquid or gas, by comparing the loss of 

 weight sustained by the same solid when weighed in pure water 

 with that sustained on its being weighed in the fluid in question. 

 According to this method, therefore, the absolute weight of a 

 body is necessary in both cases. But it is difficult to obtain this 

 unaffected by the fluid of the atmosphere in which we are im- 

 mersed. The usual operation of weighing, except where the 

 weight itself and the thing weighed happen to be equal in bulk, 

 must, from what has been said, be more or less incorrect. But 

 the specific gravity of the atmosphere being ascertained, on the 

 supposition that it is always the same, or such as to admit of its 

 changes of density being determined at all times, allowance may 

 be made for its effect on the weight of bodies, more especially 

 as it is an exceedingly light fluid, and scarcely requires to be 

 noticed, except in very nice experiments, or where the bulks of 

 bodies are very considerable. The best way of determining the 

 specific gravity of the atmosphere, and of gases generally, is to 

 weigh directly a vessel of known dimensions, when empty and 

 again when filled with the fluid in question.! It is thus found 

 that a vessel of three hundred cubic inches, for example, weighs 

 92,4 grains more when filled with air in its ordinary state, than 



t The vessel should be of considerable size, that is, sufficient to 

 contain at least three or four hundred cubic inches. It might be of 

 a globular shape, as represented in figure 212, with a narrow neck 

 and nicely fitted stop-cock C. Its capacity would be best ascertained 

 by filling it accurately with mercury, and then pouring the liquid 

 into a prismatic vessel, which might be easily measured. The air 

 might be expelled also in the same way, by connecting with the 

 neck AB a tube of about three feet in length, and suffering the 

 mercury to discharge itself through this tube, held upright with its 

 lower end immersed in the same liquid. When the mercury had 

 left the ball, upon turning the stop-cock we should effectually ex- 

 clude the air ; but there would remain a small portion of the vapour 

 of mercury. Ordinarily the air is exhausted by means of an air- 

 pump, and, although the air cannot thus be wholly withdrawn, the 

 small proportion which is left may be measured and allowance 

 made accordingly, as will be shown hereafter. 



