1837.] Self-Registering and Indestructable Barometer. 307 



^glass one. We will suppose (as is necessary) that the whole machine 

 is constructed in Europe, where every obstacle in the shape of work- 

 manship vanishes; and consider for a moment the principles of the 

 instrument. In the first place, it will be observed, that the barometer- 

 end of the balance (if we may so term it), has to support a column of 

 mercury of two square inches area, and two inches depth ; in addition 

 to which we have to add a tube of iron, whose weight is varying at 

 every fluctuation of the barometer ; for it must be recollected that the 

 part immersed in the quicksilver lessens the weight imposed upon this 

 end of the balance to the amount of the weight of an equal volume of 

 quicksilver. Let us suppose, however, that this objection is got rid of 

 (as indeed it can be), by a proportionate allowance in the divisions of 

 the scale, and proceed, in the next place, to notice what takes place 

 when the barometer rises, or when an additional weight is imposed 

 upon the barometer-end of the balance. If the centre of gravity of 

 balance coincides with the axis of support, it is evident that, on a rise 

 (however small) taking place, down goes the barometer, until the lower 

 extremity touches the bottom of the basin c — but if, as is usual in 

 practice, the centre of gravity of the balance is below the axis of sus- 

 pension, what then will be the result ? If a b represent a balance, sup- 

 ported at c, and c d, a plumb line; — on imposing an additional weight 

 at a, that end will descend until the i 

 product of the weight on a into the j 

 perpendicular a e, equals the pro- 

 duct of the t weight on the end b 

 into b f. Here we see two objec- 

 tions to the self-registering baro- 

 meter No. 1 : 1st, that, in order to 

 ascertain the amount of rise or fall, \ 

 it becomes necessary by trial to 

 adjust the axis of suspension, to 

 such height above the centre of 

 gravity, as shall render the scale 

 we may have adopted correct : 2dly, that the scale, when rendered cor- 

 rect at the zero or any other point, will not be correct for any other 

 point of indication ; unless, indeed, we here again have recourse to 

 computation, and from a table of corrections, or compute the value for 

 each division and engrave our scale accordingly. On account of fric- 

 tion (which we have no very certain means of making allowance for in 

 calculations of this sort) it is impossible to say where the centre of 

 gravity should be placed, so that an alteration in the pressure would 

 be read off from a scale of inches, as seems to be Dr. Gilchrist's inten- 

 tion; but, setting aside friction, and assuming the weight of the baro- 



