021 



BAR01IETER. 



BAROTtfETER. 



922 



led him to abandon the theory of nature's horror, though without 

 substituting any other. It has been thought that before his death he 

 suspected at least the true explanation. His pupil Torricelli first 

 imagined that the weight of the atmosphere might be the counterpoise 

 to the 32 feet of water ; or at least he was the first whom we know to 

 have applied himself to try this hypothesis by experiment. He saw 

 that, if it be a weight of air which counterpoises the 32 feet of water, it 

 must follow that by the substitution of mercury instead of water, the 

 height of the column necessary to counterpoise the weight of air would 

 be reduced in the proportion in which mercury is heavier than water. 

 For instance, that if mercury be fourteen times heavier than water, 

 bulk for bulk, the fourteenth part of 32 feet, or about 2 feet 4 inches, 

 would supply the place and produce the effect of the water. He 

 accordir gly filled a tube, more than 3 feet long, and open at one end 

 only, with mercury ; and then stopping the open end with the finger, 

 he placed the tube in an open vessel of mercury with the open end 

 downwards. On removing the finger, the mercury in the tube sank 

 until it stood in the tube at about 28 inches higher than the mercury 

 in the vessel. He thus constructed what is at this time considered the 

 best form of the barometer. 



Torriceili died shortly afterwards (1647), leaving his great discovery 

 not quite complete ; for though he had made it apparent that the 

 weight of the water and of the mercury was a counterpoise of some- 

 thing, most probably of a weight of air, the latter was not quite certain. 

 The invention however was taken up by Pascal, Mersenne, and others 

 in France, and by Boyle in England. The latter, by means of the air- 

 pump, was enabled to subject air of different degrees of density to the 

 test of the barometer. Pascal did the same ; and, in addition, first 

 suggested (in 1647) that if the mercury were sustained by the weight 

 of the air, it would necessarily fall in ascending a high mountain, by 

 the diminution of the superincumbent column of air. He accordingly 

 requested his relative, M. Perrier, to try the barometer at the summit 

 and the base of the mountain of Puy de D6me. in Auvergne, and the 

 result was that the mercury, which at the base stood 26| inches 

 (French), was only 23 inches at the summit. Pascal afterwards found 

 the same result to be sensibly shown in the ascent of a church tower and 

 of a private house. The fact was now completely established, that the 

 weight of the air upon any horizontal base was equivalent, roughly 

 speaking, to a weight of mercury of the same base, and about 28 inches 

 high. The ancient philosophers might have come to a corresponding 

 conclusion ; for, as Deluc remarks, though they had not mercurial 

 barometers, they had pumps, with which, had the taste for experi- 

 mental inquiry existed, they might easily have performed Pascal's 

 experiment. But the personification of nature answered every purpose, 

 and checked every inquiry. 



Soon after the first discovery of the barometer, many different 

 methods were imagined for improving the construction of the instru- 

 ment. The continual variations of the altitude of the mercury did not 

 escape notice ; and the idea of the weat/ter-ylass was almost contempo- 

 raneous with that of the barometer. It was observed, that changes in 

 the height of the mercury corresponded to changes of the weather, 

 though experience was not yet sufficiently extensive to decide in what 

 manner. The very gradual progress of these changes, and the frequent 

 smalluess of their amount, rendered it desirable so to construct the 

 instrument that the effect should be multiplied as much as possible. 

 And since an alteration of level in the tube of the barometer also pro- 

 duces an alteration of level in the cistern with which it communicates, 

 it soon became evident that no fixed scale of inches would serve to 

 show the difference of levels (or, as it is called, the height of the barometer) 

 merely by reading off the height of the mercury in the tube. 



We shall now give an account of the most remarkable among the 

 various constructions which have been employed or suggested. Most 

 of them are from De Luc, ' Recherches sur les Modifications de 

 1' Atmosphere. 1 In all the diagrams, a is the closed or vacuum end of 

 the tube, and p the place where the mercurial or other column com- 

 municates directly with the atmosphere. The bulbs which are drawn 

 as such should all, properly speaking, be cylinders. Enough is intro- 

 duced to show the principle of the construction, but not the method of 

 mounting the instrument. In the following descriptions each article is 

 headed by the name of the inventor, or by that of the instrument. 



Many of the following contrivances, though not at present in use, 

 may suggest ideas of value for other purposes : 



1. Torricelli. This is the simple apparatus already described. The 

 inverted tube, full of mercury, 33 or 34 inches in length, is placed in 

 the cistern of mercury. The fluid sinks until the column contained 

 between the two levels counterbalances the pressure of the air. From 

 a to there is a vacuum, named, hi honour of the inventor of the 

 instrument, the Torricellian vacuum, or at least a space only filled with 

 the vapour of mercury, which we shall presently mention. 



2 and 3. The siphon barometer, No. 2, was early adopted as a more 

 t convenient form than that of Torricelli. The pressure of the air at p 

 w counterbalanced by a column of mercury in length ir. But the 

 indications of this barometer are not nearly so great as those of 

 Torricelli's ; for an inch of variation in the difference of levels makes 

 the mercury in the closed tube descend half an inch, and that in the 

 open tube ascend half an inch, or vice versa ; thus altering s v by one 

 inch (a falling half an inch, and p rising half an inch). In Torricelli's 



by the tube excluded) were twenty times that of the tube, then a 

 diminution of an inch in s v would be marked by a fall of |'i of an 



inch in the tube, and a rise of ^ of an inch in the cistern ; for the 

 mercury which is driven out of the tube causes a little addition to the 

 cylinder of mercury in the cistern, of twenty times the base it occupied 

 while in the tube, and therefore of only one-twentieth of its height. 

 No. 3 is a siphon barometer, with a similar method of increasing the 

 variation in the tube. The siphon terminates in a basin of greater 

 diameter than the tube. If the horizontal section of the basin be 

 twenty times that of the tube, we have again the case just explained. 



4. Descartes. Here we have the top of a barometer so constructed, 

 that a narrow tube shall open into a wider cistern, which opens down- 

 wards into a tube. Any light fluid, say oil, is first poured in, and 

 afterwards mercury ; the vacuum is then made as in Torricelli's expe- 

 riment, and the quantity of oil and the diameter of the cistern are so 

 adjusted, that the extreme variations of the weight of the atmosphere 

 shall allow some mercury to remain in the cistern. Let us say that the 

 specific gravity of the oil is one-twentieth that of mercury, or that a 

 column of oil is of the same weight as a column of mercury one- 

 twentieth of its length ; and let us suppose a fall of an inch in the 

 purely mercurial barometer. Let us also suppose the horizontal 



barometer, if the horizontal section of the cistern (the part occupied section of the cylinder to be ten times that of the tube above and 



