Admiral FitzRoy writes:--"The words on the scales of barometers should not be so much regarded, for weather indications, as the rising or falling of the mercury; for if it stands at _Changeable_, and then rises a little towards _Fair_, it presages a change of wind or weather, though not so great as if the mercury had risen higher; and, on the contrary, if the mercury stands above _Fair_ and falls, it presages a change, though not to so great a degree as if it had stood lower; besides which, the direction and force of wind are not in any way noticed. It is not from the point at which the mercury stands that we are alone to form a judgment of the state of the weather, but from its _rising_ or _falling_; and from the movements of immediately preceding days as well as hours, keeping in mind effects of change of _direction_ and dryness, or moisture, as well as alteration of force or strength of wind."[1]
=13. Correction due to Capacity of Cistern.=--These barometers, having no adjustment for the zero of the scale, require a correction for the varying level of the mercury in the cistern, when the observations are required for strict comparison with other barometric observations, or when they are registered for scientific purposes; but for the common purpose of predicting the weather, this correction is unnecessary. The neutral point, and the ratio of the bore of the tube to the diameter of the cistern, must be known (see p. 3). Then the capacity correction, as it is termed, is found as follows:--Take the fractional part, expressed by the capacity ratio, of the difference between the observed reading and the height of the neutral point; then, if the mercury stand _below_ the neutral point, _subtract_ this result from the reading; if it stand _above_, _add_ it to the reading.
For example, suppose the neutral point to be 29·95 inches, and the capacity ratio 1/50, required the correction when the barometer reads 30·78.
Here 30·78 - 29·95 = 0·83 Correction = 0·83/50 = +0·02 nearly. Scale reading 30·78 ----- Correct reading 30·80 =====
Of course the correction could as easily be found to three decimal places, if desirable. It is evident that the correction is more important the greater the distance of the top of the mercury from the neutral point.