1823.] the Bai'Ometer to the Measurement of Heights * 107 



proposition of M. Daubuisson deserves to be received. Table, 

 No. 4, renders this correction extremely easy : the whole opera- 

 tion is reduced to diminishing by a tenth the number which the 

 table gives for mercury. 



The correction for the dilatation of brass always diminishes 

 the correction for the dilatation of mercury, and even in the case 

 where we reduce the temperature of the warmer to that of the 

 colder barometer. In fact, the pyrometric variation of the scale 

 proceeds in an opposite direction to that of mercury. If heat 

 lengthen the scale, the column of mercury is measured by a less 

 number of divisions : if cold contract it, by a greater. In the 

 first case, the quantity to be subtracted from the height of the 

 mercury is diminished by the quantity to be added to the length 

 of the scale. In the second case, the quantity to be added to 

 the height of the mercury is diminished by that which is to be 

 subtracted from the length of the scale. This rule must not be 

 forgotten, if we adopt the plan of reducing all observations to a 

 constant temperature, as, for example, to 12*5°. 



In respect to using these tables, it may be observed, that it is 

 altogether indifferent by what scale the barometer is divided, 

 provided its lesser divisions be always expressed in decimal parts 

 of the integers. 



It is of consequence to the accuracy of barometrical measure-' 

 ments, to have barometers in which the mercury stands at its 

 real and absolute elevation. In instruments of the cistern con- 

 struction, this is never the case. The column is depressed 

 owing to the force of capillarity, which is the more considerable 

 in proportion as the diameter of the tube is less : it exceeds 

 •039 inch in tubes whose interior diameter is from '19 to '20 

 inch. M. de Laplace has calculated a table of these depressions 

 which is here given. In order to use it, we must measure accu- 

 rately the interior diameter of the tubes we employ, and add to 

 the height of the mercury, the quantity in the table, answering 

 to the number nearest the given diameter. 



The method of calculation just explained, although very expe- 

 ditious, maybe, perhaps, somewhat less so than methods founded 

 on tables specially adapted for giving the result at once, and 

 particularly if such tables are carried to such an extent as to 

 dispense with all interpolation ; but it will always retain the 

 advantacre of a greater generality, and it appears to me to recom- 

 mend itself not only by the exactness and faciHty with which it 

 is performed, but above all by the convenience which it possesses 

 of being equally suited to calculators of all countries ; and of 

 admitting any alteration which it may be judged necessary to 

 make in the constant coefficient. 



These considerations, however, diminish nothing of the merit 

 of many very ingenious methods which have been substituted for 

 the logarithmic; but of these the most remarkable for brevity 

 have necessarily the fault of leaving to the charge of the 



