METEOROLOGY. 475 



Bctcrmlmdlon of the height of the Silla dc Caracas, at noon March 26, 1867. 

 At lower station (wliarf at Laguayra) : 



Barometer 7o9'"™.32, or 29.904 English inches. 



Attached thermometer 24°.2 C, or 75°. .56 Fah. 



Thermometer in open air 24°.2 C, or 75°.56 Fah. 

 At njjper station, (top of the Silla do Caracas,) in latitude 10° SI' 15" north : 



Barometer 557'""\5, or 21.956 English inches. 



Attached harometer 13°.S C, or 54°.84 Fah. 



Thermometer in open air 13°. 8 C, or 54°.84 Fah. 

 Hence by Baily's tables, elevation of the Silla de Caracas 8,658 English feet, 

 or 5734.5 feet above Caracas. 



APPENDIX A. 



[From the Reader, Dec. ]0, 1864, p. 740.] 



It appears to be the ordinary practice of instrument makers, when construct- 

 ing a barometer with the English and French scales, to turn to the tables for 

 the conversion of inches into millimetres, or vice versa, and to assume that the 

 equivalents there found are to be implicitly adopted. I am far from blaming 

 them for this assumption, but merely Avish to show that it should in future cease 

 to be acted upon. An examination into the data on which these tables are con- 

 structed shows that they merely profess to give the equivalent of English inches 

 at the standard temperature of the yard (62° F.) in millimetres at the standard 

 tenqyerature of the metre (0° C., or 32° F.) Hence the reading of the metiical 

 scale of a barometer corresponding to any given number of inches should be the 

 tabular equivalent less a coiTection for the expansion of the scale between the 

 respective standard temperatures (0° C, 62° F.,) or for a range of 30° F. (16°.67 

 C.) It will be at once apparent that, at the same pressure, the amount of this 

 correction will be constant, whatever be the temperature common to the two sccdes. 



An example will render my meaning clearer. Let the barometric reading on 

 the English scale be 31 inches, the equivalent of which in the tables is 787.37 

 millimetres, based on the assumption that the temperatures of the scales are respect- 

 ively 62° and 32° F. If the attached thennometer indicates 62° it is clear that 

 the requisite condition is not realized in the case of the metric scale, which must 

 therefore be corrected. Let A be the linear expansion of brass for 1° C. = 

 0.000018782 ; B the metrical reading = 787.39'^'", and t the temperature = 62° 

 F., or 76°.67 C. ; then 



A B ^ = 0.000018782 times 787.39 times 16.67 = 0.247™°^ the amount to 

 be deducted from the reading of the metrical scale equivalent to 31 inches, as 

 given in the table, in order to reduce it to the same temperature as the English 

 scale. 



Unless this be done a discrepancy must always become apparent in the 

 reduced readings of the two scales. For instance, in the case of a mountain 

 barometer, by one of our leading makers, and now in my possession, I find 

 761.99'"'" are made to correspond (following the authority of the tables) to 30 

 inches. If the temperature of the attached thermometer l)e 62° F., the respect- 

 ive readings reduced to the freezing point become 30 — 0.090 = 29.910 inches 

 at the standard temperature, (62° F.,) 761.99 — 2.05 = 759.94""^' at the stand- 

 ard temperature 0° C, (32° F.) Turning to the tables for the comparison of the 

 scales, we lind that 29.910'" at 62° F. are equivalent to 759.70'""' at 0° C, 

 instead of 759.94 as above. If the barometer was properly constructed we 

 ought to have now 761.75""", as corresponding to 30 inches, and the figui^es 



