C'jitiputatwTt, and Lhi-ts ef Errcr. ,27 



To render this more e^y by an example, fuppofe the height of the mercury in the baro- 

 meter to be 78 centimetres, .nd the inftrument being empty to be plunged in the mercury 

 to the poun C. It IS then covered, and raifed until the fmall column of mercury DE is 

 fufpended, for example, at the height of fix centimetres. The Internal air, which was at 

 firft comprelTed by a force reprefenteJ by 78 centimetres, is now compreffed only by a 

 force reprefented by 78 —6, or 72 centimetres. 



Suppofe it to be obfervcd at the fame time, by means of the graduations of the fecond 

 fcale, that the capacity of the part CD of the tube which the mercury has quitted is 2 cubic 

 centimetres. Then by the rule V X 2 give 24 cubical centimetres, which is the volume of 

 the air included in the inftrument when the mercury rofe as high as C in the tube. 



The body of which the volume is to be afcertained, muft then be placed in the capfule, 

 and the operation repeated. Suppofe, in this cafe, the column of mercury fufpended to be 

 eight centimetres, when the capacity of the part CD of the tube is equal to 2 centimetres 

 cube. Then the greateft preflure being denoted by 78 centimetres, as before, the leaft will 

 be 70 cemimetres, the difference of the prefTurcs being 8, and the difference of the volumes 

 two cubical centimetres. Hence S x i gives the bulk of the included air under the great- 

 eft preffure 17,5 cubic centimetres. If therefore 17,5 centimetres be taken from 24 cen- 

 tlir.etres, or the cipacityof the inftrument when empty, the difi-erence 6,5 cubic centimetres 

 will exprefs the volume of the body which was introduced. And if the abfolute weight of 

 the body be multiplied by its bulk in centimetres, and divided by the abfolute weight of one 

 cubic centimetre of diftilled water, the quotient will exprefs the fpecific gravity of the body 

 in the common form of the tables where diftilled water is taken as unity, or the term of com . 

 parifon. 



After this defcription and explanation of the ufe of his inftrument, the author proceeds 

 with the candour and acutenefs of a philofopher to afcertain the limits of error in the refults; 

 an objea feldora fufficiently attended to in the inveftigation of natural phenomena. From his 

 refults it appears, that with the dimenfions he has affumed, and the method prefcribed for 

 operating, the errors may affeft the fecond figure. He likewife gives the formula by means 

 of wliich the inftrument itfelf may be made to fupply the want of a barometer in afcertain- 

 ing the greateft prelTure. He likewife adverts to the errors which maybe produced by change 

 of temperature. To prevent thcfe as much as poffible, the adual form of the Inftrument 

 and arrangements of its auxiliary parts are fettled as in Fig. 3, by which means the approach 

 of the hand near the vefl'el and its tube is avoided. In this figure the vertical poCtiun of 

 the tube is fecured by the fufpenlion of the veflcl, and a perforation in the table through 

 which the tube palTes. The table itfelf fupports the capfule in its firft pofitlon, namely 

 that at which the cover is required to be put on. 



It feems probable, from the author's general rcdeaions Immediately preceding the expla- 

 nation of the figures, that he had not entirely finilhed his meditations 011 this fubjed. If 

 he had, I think it likely that he would have determined his prcftures as well as the niea- 

 furcs of bulks by weight. For it may be eafily underftood, that if the whole inftrument 

 were fct to its pofitions by fufpending it to one arm of a balance at H Fig. 3, the quantity 

 of countcrpoife when in cquilibrio might be applied to determine the prclTurcs to a diiTCc 

 of accuracy much greater than can be obtained by linear mcafurcment. 



5 VIII. U/.f,,i 



