BAR 



many ten liundredth parts of an inch, and numbfred nc- 

 cordiEgly, lo, 20, 30, 40, &c. parts of an inch; then the 

 Vernier gives the unit to each ten, tlius : fct the index very 

 accurately to the top of tlie furface of the mercury E ; and 

 if at the fa:r.e time, the beginning of the divilions at C co- 

 incide with a line of divifion in the fcale B, then it fhews 

 the altitude of the mercury in inches and tenths of an inch 

 exaAly. But if the index line C of the Vernier fall between 

 two divilions or tenths on the fcale B, then there will be a 

 coincidence of lines in both at that n\imber of the Vernier, 

 which (hews liow many tenth parts of that tenth the index 

 of the Vernier has palTed the lall decimal divifion of the 

 fcale. E. G. Suppofe the index of the Vernier were to 

 point fomcwhere between the fixth and feventh tenth above 

 30 on the fcale ; then, if by looking down the Vernier, you 

 obferve the coincidence at number 8, this lliews that the al- 

 titude of the mercury is 30 inches, and. 68 parts of a hun- 

 dredth of another inch, or fimply thus, 30,68 inches. See 

 Vernier. The fcrew/, in_//f. 84. ferves to prtfs the mer- 

 cury up into the tube, when the inftrnment is to be moved, 

 and thus to render the inftrument z pji-tablc barometer. 



The barometer belonging to the honfe of the Royal So- 

 ciety is of the cillern kind ; and the Hon. Mr. Cavendilh 

 prefers this form to that of the fyphon kind, becaufe both 

 the trouble of obferving and the error of obfervation are 

 lefs, as in the latter we are liable to an error in obferving 

 both legs. Moreover he remarks, that the quickfilver can 

 hardly fail of fettling more exaftly in the former than in the 

 latter ; for the error in the fettling of the quickfilver can 

 proceed only from the adhefion of its edge to the fides of 

 the tube. In the latter the adhefion may take place in two 

 legs, but in the former only in one ; and befides, as the air 

 has neceffarily accefs to the lower leg of the fyphon baro- 

 meter, the adhefion of the quickfilver in it to the tube will 

 moil probably be different according to the degree of dry- 

 nefs or cleannefs of the glafs. It is true, as M. De Luc ob- 

 ferves, that the ciftern barometer does not give the true 

 preffure of the atmofphere ; the quickfilver in it being a 

 little depreffed on the fame principle as in capillary tubes. 

 But it appears by calculation, that in the barometer of the 

 fociety, the error arifing from the alteration of the height 

 of the quickfilver in the ciftern can fcarcely ever amount to 

 fo much as -s-Jiith of an inch. In this barometer, the 

 height of the quickfilver is eftimated by the top of its con- 

 vex furface, and not by the edge where it touches the 

 glafs ; the index being properly adapted for that purpofe ; 

 and this manner of obterving is more accurate than the other. 

 Phil. Tranf. vol. Ixvi. p. 38 1. 



As foon as it was difcovered that the different heights of 

 the mercury indicated by the barometer were in fome degree 

 connefted with the ftate of the weather, and that it might 

 be applied to the purpofe of a " weather-glafs," many at- 

 tempts were made to render the changes in it more fen- 

 fible, and fo to meafnre the variations of the weight of 

 the atmofphere more accurately ; and thefe attempts have 

 given rife to a great number of barometers of different 

 Itrudures, deviating from the fimplicity of the common ba- 

 rometer, and at the fame time lefs accurate. Hence the 

 wheel barometer, diagonal barometer, horizontal barometer, 

 pendant barometer, &c. 



Dcs Cartes fuggefted the firft; method of increafing the 

 apparent fenfibility, or enlarging the fcale of variation, of 

 the barometer, though he did not live to execute it. He pro- 

 pofed a tube AB {Plate IX. Pneumatics, ^g. 77.) about twen- 

 ty-feven inches long, terminating in a cylindric veffel CD : 

 one half of which veffel, connefted above with a long tube 

 of a very fmall bore, feakd Kt top, and exhaufled of Us air, 



BAR 



was to be filled with water extending up into the fnr.all tube ; 

 the other part of the veffel, and the lower part of the tube, 

 were to be filled with meicury. Whenever the mercury rofe 

 in the cylinder, it would force up a proportional quantity 

 of water into the narrow tube, where it would have a 

 confiderably larger range than that of the mercury in 

 the cylinder: neglefting the weight or prefTurc of the 

 water, the m.otion of the water and of the mercury 

 would be in the invcrfe ratio of the iquarcs of the diameters 

 of the vefi'els containing them. But the water preffes on 

 the mercury according to its height ; and therefore if the 

 whole range of the mercury in the cylinder, or in a common 

 barometer, were fuppofed to be two inches, tlie fpecific gra- 

 vity of water to that of mercurj' as r to 14, and the dif- 

 ference between the diameters of the cylinder and tube a 

 maximum or infinite, then the entire fcale of variation in 

 this inftrument would be twenty-eight inches ; or the ex- 

 tent of this fcale would be to that of the common barome- 

 ter in the inverfe ratio of the fpecific gravity of water to 

 that of mercury. It is evident that in pradlice it would be 

 fomewhat lefs than twenty-eight inches. Huygens con- 

 ftrufted a barometer of this kind ; but here, though the 

 column fufpended was larger, and confequently the variation 

 greater, yet the air imprifoned in the water getting loofe 

 by degrees, filled the void fpace in the top, and fo ruined 

 the machine. 



Huygens then thought of changing the conftruftionof the 

 barometer, and of placing the mercury at top, and the water 

 at bottom, in the following manner: ADG (/g. 78.) is a 

 bent tube hermetically fealed in A, and open in G ; the 

 cylindric veflels BC and FE are equal, and about twenty- 

 nine inches apart ; the diameter of the tube is about a line, 

 that of each veffel fifteen lines, and the depth of the vef- 

 fels is about ten ; the tube is filled with mercury (the com- 

 mon barometer ftanding about twenty-nine inches) which 

 will be fufpended between the middle of the veffel FE, and 

 that of the veffel BC ; the remaining fpace to A being 

 void both of mercury and air ; laftly, common water, tinged 

 with a fixth part of ngi/a regis to prevent its freezing, is 

 poured into the tube FG till it rifes a foot above the mer- 

 cury in DF. 



When the mercury rifing above the level of that con- 

 tained in FE, through the tube AD, becomes a balance to 

 the weight of the atmofphere ; as the atmofphere increafes, 

 the column of mercury will increafe, confequently the water 

 will defcend ; as the atmofphere again grows lighter, the 

 column of mercury will defcend, and the water afcend. This 

 double barometer, as it was called, which is nearly the fame 

 with that of Dr. Hooke, will therefore difcover much minuter 

 alterations in the air than the common one ; for, inftead of 

 two inches, the fluid will here vary tvi'o feet ; and by en- 

 larging the diameters of the cylinders, that variation may 

 be ftill increafed ; but it has this inconvenience, befides 

 others, that the water will evaporate, and fo render the al- 

 terations precarious ; though the evaporations be in fome 

 meafure prevented by a drop of oil of fweet almonds fwim- 

 ming at top : the column of water will likewife be fenfibly 

 affefted by heat and cold. 



The double barometer of Dr. Hooke was invented in the 

 year 1668, and is defcribed in the Phil. Tranf. N° 185. 

 The invention was claimed by Huygens and De la Hire ; 

 but it fufficicntly appears, that Hooke was the original in- 

 ventor. (See De Luc's Recherches, vol. i. p. 18.) This 

 confifts of a compound tube ABCDEFG (f-s-19-)i of 

 which the parts AB and DE are equally wide, and EFG 

 as much nairower as it is propofed to enlarge the fcale. 

 The parts AB and EG are made as cylindrical as poffible. 



The 



