17 



In consequence of this the mercury column is actually depressed a 

 slight amount and never indicates the true barometric height. This 

 source of error is one of the most troublesome to which barometers 

 are subject, as the capillarity is never quite constant and there is no 

 practicable method by which its changing value can be accurately 

 determined in the daily use of an instrument. 



The error due to capillarity is nearly always eliminated as far as 

 possible from the scale reading by adjusting the scale so that allow- 

 ance will be made for the average capillary depression. If an ordi- 

 nary barometer be carefully examined it will be found that the 30- 

 inch mark on the scale is appreciably less than 30 inches from the 

 ivory point. In general, the difference represents the amount the 

 mercurial column is depressed by capillarity. A portion of a barom- 

 eter scale is shown enlarged at v, Figure 2 (at the top and at the 

 right). The index line v is made accurately 30 inches from the 

 ivory point, but the 30-inch line on the scale is shown set slightly be- 

 low to offset the capillary depression. 



(2) Correction for imperfect vacviuvi. — It is generally assumed 

 that the space in a barometer tube above the mercurial column is a 

 perfect vacuum, and that there is no downward pressure upon the 

 top of the column of mercury. This, however, is not strictly the case 

 in any instance, and often an appreciable quantity of air or water 

 vapor is present. Any vapor that the mercury may give off is, of 

 course, always present. This latter, however, is very small and is 

 never considered except in the most refined investigations. If, there- 

 fore, any such pressures exist upon the top of the column it will be de- 

 pressed, and a correction, which may be properly called correction 

 for imperfect vacuum or reduction to perfect vacuum, should be ap- 

 plied. Such a correction will vary with both the temperature and 

 the volume of the space. If the trace of air present is slight, as is 

 nearly always the case in any good barometer, the correction for 

 vacuum will be nearly constant, provided the volume is not changed 

 much by great changes of pressure, as the changes corresponding to 

 ordinary changes in temperature are comparatively small. There- 

 fore, in ordinary observatory barometers this correction, like the one 

 for capillarity, is included in the correction for instrumental error. 

 When, however, a barometer is used at both high and low pressures, 

 the volume of the vacuum space may change many fold, and in such 

 a case any error due to imperfect vacuum is far from being constant. 



(3) Connection for instru/mental or scale error. — Errors arising 

 from several independent sources are embraced under this designa- 

 tion, as, for example : (a) The graduated scale may not be adjusted so 

 perfectly that its divisions are at exactly the right distance from the 

 ivory point; (6) the sighting edge of the vernier may not be true or 

 in proper correspondence with its zero graduation line; (c) the 

 imavoidable errors and irregularities in the graduations of the barom- 

 eter scale itself also introduce different errors from point to point 

 along the scale. Nevertheless, sufficient precision in scale gradua- 

 tion is easily attainable even in a scale that is only fairly good, and 

 such errors are generally so small as to be unimportant in ordinary 

 barometric observations and are seldom considered. 



As stated in a previous paragraph, the combined effect of such 

 sources of error as those just mentioned gives rise to what is generally 

 called the correction for instrumental or scale error. 



