Notes on Barometer CoTistruction. 37 



greater than that of the barometer tube to be calibrated, 

 must be used. Immediately under the bowl of this funnel 

 is a stopcock which, when the point of this long funnel tube 

 is lowered to the bottom of the barometer tube, enables us 

 to regulate the supply of mercury, so that the surface of the 

 fluid mercury rises slowly and equably, filling the tube 

 without locking in a single bubble of air against the 

 inner glass surface of the barometer tube. The glass 

 measure fitted to the lower end of the barometer tube, as 

 already described, is a spheroid with tubular ends. There 

 is a narrow vertical glass tube forming its upper opening, 

 and on this narrow glass tube a measuring mark is made ; 

 a second mark is also placed on the tube below the lower 

 orifice of the bulb. With this arrangement we can calibrate 

 the barometer tube. We first fill the tube under trial with 

 mercury ; we then open the stopcock of supply and allow 

 mercury to run oif until it has reached the trait x below the 

 bulb. We now mark on the barometer tube the position of 

 the upper surface of its mercurial column. We next open 

 the stopcock of supply, until we have filled the measuring 

 bulb to its upper mark y, when we mark the level to which 

 the upper surface of the mercury has descended in the 

 barometer tube. The supply cock being shut ofij we next 

 open the discharge cock, allowing mercury to flow slowly 

 out until the lower mark is reached. In this way the 

 measuring bulb is slowly and accurately alternately filled 

 and emptied between the two gauge marks, and after each 

 filling the level of the mercury in the barometer tube is 

 carefully registered on it. This is continued until the 

 barometer tube is almost or quite emptied, by which time 

 we have marked it with subdivisions throughout its length, 

 each of which we know to be of capacity equal to the rest, 

 and from their, several distances apart the diameter of every 

 portion of the tube can be computed. The temperature of 

 the mercury and the weight of the bulb measure of mercury 

 should be noted, and when extreme accuracy is the aim 

 there are other influences to consider and allow for ; but 

 the modus operandi is essentially what I have described 

 whenever a barometer tube, or indeed a straight glass tube 

 of any kind, is to be calibrated. The data for correcting 

 the bulk of the mercury for temperature, &c., &c., are fully 

 set forth in physical treatises,, and therefore I need not 

 further allude to them in this place. 



