1838.] Description of Sketches of a Self-Registering Barometer, 125 



tity of mercury, but the greatest disadvantage of it is, that, unless the 

 cistern in which the whole floats is very wide, access to the bottom of 

 the tube, to avoid the entrance of air, when it is wished to dismantle the 

 instrument, with a view to conveyance, &c., is prevented. It would not 

 be difficult to contrive a method for overcoming this objection. It has, 

 however, occurred to me, that the float may be placed in a cistern abcve 

 the tube, so that the tube acts as ballast to itself, and the superiority of 

 this plan over the former will at once appear. 



i?'^^. I A is the pedestal, furnished with screws and spirit levels, 

 ««, to adjust the instrument to the perpendicular. BB pillars to which 

 the whole of the apparatus is attached. C is a cistern, consisting of a 

 broad upper portion b, and a narrow lower portion c ; D is a second cis- 

 tern of similiar construction. In C is afloat to which the rod rising 

 perpendicularly out of it is attached. The other rods connecting the 

 float and the tube E can easily be traced. FFFF are friction rollers 

 which run in grooves in the pillars, and are intended to confine the 

 tube to one line of ascent and descent. The necessity for friction rol- 

 lers may at first sight appear an objection to the instrument, but when 

 it is remembered that the only power so capable of moving the tube 

 from the perpendicular is the delicate spring which keeps the pencil 

 against the register roller, this objection will vanish. H is the regis- 

 ter roller, &c. described in a previous communication. I is a scale and 

 vernier index, which will show the height of the mercury at any given 

 time without the necessity of disturbing the registering operation— in. 

 spection of the sketch will show that this can be raised or lowered on 

 the pillar ; and the tangent screw d at top admits of its minute adjust- 

 ment. The cistern D can also be adapted to any altitude of station by 

 means of the thumb screws near pillars and the screw c working in the 

 collar F. By this means the top surface of mercury may be brought to 

 the middle of the expanded portion of the tube when the instrument is 

 put in operation. It is essential in this variety of instrument that the 

 cisterns C and D be furnished with expanded tops, because on any 

 change of atmospheric pressure, the relative height of surface of mer- 

 cury in those cisterns varies — in the case of C from the dip or immer- 

 sion of the float, and in the case of D from the same with reference to 

 the end of the tube ; also from mercury entering or leaving it. The 

 principle observed here is that of " capaciti/^ in the mountain baro- 

 meter. 



