OF ARTS AND SCIENCES. 243 



For testing mercurial thermometers in the vapor of substances boil- 

 ing at high temperatures, the following apparatus has been employed. 

 It is similar to the boiler of the larger apparatus. Into the bottom of 

 a thin brass tube, of about 5.3 cm. diameter and 20 cm. high, is brazed 

 a thicker plate. Within this tube stands a shorter tube of about 4.5 cm. 

 diameter notched at the bottom edge, and having a somewhat eccentric 

 collar at about 2 cm. from the bottom, to hold it in place and prevent 

 the vapor from freely rising into the outer jacket. The cover fits with 

 a flange into the top of the outer tube and is split along a diameter. 

 The boring for the insertion of the thermometer is in the centre. 

 Through one half of the cover passes a thin tube about 60 cm. long 

 and 1 cm. internal diameter, projecting about 15 cm. below the cover. 

 This lower portion thus extends nearly to the surface of the liquid in 

 the outer jacket, being placed of course in the larger side of that 

 jacket, and serves as an escape or condenser tube. 



Pressure Regulator. 



This has been elsewhere described in full,* and is shown in the 

 figure. A Richards water-jet aspirator drew air from a b c, and c was 

 connected with the apparatus to be exhausted. The - 



small glass tube efp was drawn out to an open point «= 



at p. As the exhaustion proceeded, the mercury rose « 



in the larger glass tube/ and in f, until the level in the 

 open cistern h g fell below p, whereupon the mercury in 

 /would flow over into j, followed by a sudden inrush of 

 air, thus increasing the internal air pressure and causing 

 the mercury to fall somewhat iaj and rise at h g, thus 

 closing p. Repetitions of this process would occur until 

 a steady condition was reached, when a nearly regular 

 stream of air and mercury globules would flow con- 

 tinuously through/. To maintain steady action, proper 

 relative sizes of tubes and openings must be discovered, 



and some constriction should be placed in ij, and a vessel of large 

 capacity should be present in the circuit. The amount of exhaustion 

 can of course be regulated by the quantity of mercury in the cistern, 

 and by the lengths of the tubes. The pressures were of course 

 measured by a separate mercury column and the barometer. 



* Proc. Amer. Acad , xxi. 1 (1885) ; Technology Quarterly, i. (1886). 



