36 Notes on Barometer Construction. 



zinc. Or we may choose to be more exact, and properly 

 calibrate our tube throughout ; although it must be here 

 admitted that even for a syphon barometer it is only a few 

 inches of each end of the tube which is required to be of 

 uniform diameter. For calibration, if the interior diameter 

 of the tube be small, say not exceeding two, or at the most 

 three millimetres, we may pass a cylinder of mercury of 

 known weight from end to end of the tube, accurately 

 measuring the length of this thread of mercury progressively 

 during its course ; this will give data from which we may 

 calculate the mean diameter of the bore of the tube in every 

 portion of its length.* 



For the calibration of wide tubes we may close one end, 

 and, fixing the tube in a vertical position, weigh or accu- 

 rately measure into it definite constant quantities of mer- 

 cury. Or a method well calculated to avoid air bubbles 

 may be practised by fitting the lower closed end of the tube 

 with a glass reservoir, furnished with tubular terminations 

 and glass or steel stopcocks. This reservoir with its tubes 

 has the form of the letter U, the reservoir forming the thick 

 arm of the letter (see Fig. 1). The parallel vertical tubular 

 branch representing the thin arm contains a stopcock of 

 supply, while a second stopcock for discharge of the mercury 

 from the reservoir is placed at the lower portion or bend of 

 the U. The whole requires to be fixed on a vertical board, 

 and a funnel with capillary lower termination, of a length 



* For purposes for which it is convenient to gauge, with a metallic gauge, 

 the interior diameter of the two ends of the glass tube, the calculation for 

 the estimation of the relative diameter of the intermediate parts becomes 

 very simple, as the following example chosen as affording a simple illustra- 

 tion will show : — Say diameter at each or either end is found by the gauge 

 to be 4 millimetres, and that we introduce a cylinder of mercury measuring 

 in this part of the tube 10 millimetres in length. Suppose that we pass this 

 column along towards the centre of the tube to a position in which its 

 length is exactly doubled, becoming 20 millimetres, the cubic measurement 

 of the mercury is 4* : '7854 : 10 ; but for our purpose, as the proportion '7854 

 to unity is common to all the sectional areas we may discard this factor '7854, 

 and thus we deal with 4* : 10 = 160. This in the portion of the tube where 

 the length of the mercurial cylinder is doubled, occupying 20 millimetres, 

 divided by the latter (i^V = 8) will give a quotient of 8, the square root of 

 which, say 2*84 millimetres, is the diameter of the centre of this portion of 

 the tube ; and so indeed for any other part, the square root of the quotient 

 obtained by dividing 160 by the length of the mercurial column in that part 

 will give the local diameter. Of course in tubes selected for their apparently 

 near approach to a perfectly cylindrical form the length of the mercurial 

 calibrating column will be nearly uniform throughout, but whatever dif- 

 ferences there may be are calculable from results obtained by the method 

 described. See illustration A. 



i 



