350 Mr. A. Bird on a Water-Barometer.- 



vent the cement from running into and stopping up the compo 

 tube. The glass tube C, perfectly clean inside, was now placed 

 in the socket; and being most carefully steadied to keep it 

 upright, six inches of dry sand were poured down to keep the 

 cement from rising up the glass tube C. 



The cement was composed of two parts of gutta percha and 

 one part of common black pitch. These two substances were 

 heated in an iron ladle with a lip, till they became perfectly fluid 

 and quite free from froth. A "copper bit" used by plumbers 

 having been heated to low soldering-heat, a small quantity of 

 the cement was poured into the socket. The copper bit was 

 then applied to the outside, the effect being to perfectly liquefy 

 the cement in situ. A little more of the hot cement was then 

 poured in, and again the heated copper hit was applied till 

 the socket was quite full of very fluid cement without any air- 

 cavities therein. As the cement cooled, it clung to the glass 

 and metal, and became absolutely solid and air-tight. If the 

 cement is poured in all at once, it is impossible to prevent cre- 

 vices, which will let in air when the barometer is filled, causing 

 the water gradually to descend till it falls out of the instrument. 



A place being chosen on the staircase of my house, a flat 

 board, 7 feet long and 1 foot wide, was fastened to the wall, 

 upon which board was fixed the socketed glass tube C, and gra- 

 duated scale F, from the top of which 422 inches were most care- 

 fully measured down to the "zero "-point E beside the cistern. 



The scale F is to the right of the glass tube. It is made 

 of well-seasoned boxwood, and is graduated to inches and 

 tenths. The sliding-tube G, with the vernier H, is between the 

 glass tube and the boxwood scale F. On the left side of the 

 glass tube C is another sliding-tube g, with a vernier h, to 

 record position of top of tidal column of water at 9 a.m. the 

 morning previously. 



The glass tube, scale, and verniers having been securely 

 placed on the board and perfectly upright, the gas-fitter proceeded 

 to connect, by soldering, the remainder of the compo tube 

 above the glass tube C, which was continued upwards till it 

 entered nearly at the bottom into a round vessel K, made of 

 zinc, 4 inches in diameter and 18 inches high. Inside the vessel 

 the tube coils round in a spiral, like the worm of a still. This 

 vessel and spiral are not necessary to the action of the baro- 

 meter; but as the spiral is in the part of the tube in which is 

 the vacuum- chamber, it gives the opportunity of artificially 

 cooling with ice or snow the included aqueous vapour, and thus 

 determining by actual experiment the amount of correction 

 required. 



If the experiment of cooling the included vapour to 32° be 



