336 Mr Forbes on the Defects of the Sympiesometer, 



poses a cylinder of mercury, the most dense fluid discovered, 

 of perhaps an inch-Jong and \ in diameter. This we appre- 

 hend will be readily admitted ; and it is equally incontroverti- 

 ble, that the accuracy of the sympiesometer fundamentally de- 

 pends on the precision of the correction for changes of tempe- 

 rature, elegantly performed by a sliding scale. Whence, if 

 the atmosphere be in a variable condition, and we shall find 

 that at almost all times it is sufficiently so to produce an ef- 

 fect, the thermometer, from its less sensibility than the ma- 

 gazine of hydrogen, which is the important part of the instru- 

 ment, (the oil acting merely as an index), will show a tempera- 

 ture more or less different from that required for the true cor- 

 rection, which will therefore be erroneously made. Suppose 

 the temperature of the air sinking, (for in the case of heights 

 the instrument, at one station at least, may be presumed to be 

 in the open air,) the instrument itself will have parted with a 

 minute quantity of caloric before the thermometer. The scale 

 therefore depending on the latter will be pushed lower down 

 than if the true temperature had been indicated, wherefore 

 the oil will stand opposite a higher point on the sliding scale, 

 which is divided into inches of mercury,- and a logarithimic 

 line of fathoms, than indicates the actual pressure of the atmo- 

 sphere. Conversely, if the temperature be rising, the pres- 

 sure indicated will be too small. 



If this source of error was ever thought of, probably it was 

 considered too minute to be sensible. We must therefore en- 

 deavour to substantiate it by experiments, in detailing which, 

 I would premise, that, as far as they go, I put perfect confi- 

 dence in their accuracy, though I could wish some additional 

 observations had been made, both under circumstances consi- 

 derably different, and in repetition of the observations on re- 

 turning to the first station ; though the latter precaution we 

 shall find will little affect the general conclusions. 



The upper station, which we shall designate by A, was the 

 room in which the sympiesometer usually hung in Colinton 

 House, at 399 feet above the mean level of the sea. The 

 lower one, B, was on the bank of the Water of Leith, in a 

 deep valley, which offered a very fit locality for barometric ex- 



